dask-collections.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8b598cf5-4995-4c02-9237-03b38848bb8f",
   "metadata": {},
   "source": [
    "<img src=\"assets/dask_horizontal.svg\"\n",
    "     width=\"45%\"\n",
    "     alt=\"Dask logo\\\">\n",
    "     \n",
    "# Parallel Computing in Python with Dask\n",
    "\n",
    "This notebook provides a high-level overview of Dask. We discuss why you might want to use Dask, the high-level and low-level APIs for generating computational graphs, and the schedulers which allow for the parallel execution of these graphs."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "305e71dc-d11c-4e87-946c-cf7152634128",
   "metadata": {},
   "source": [
    "# Overview\n",
    "\n",
    "### What is Dask?\n",
    "\n",
    "- Dask is a flexible, open source library for parallel computing in Python\n",
    "\n",
    "    - GitHub: [https://github.com/dask/dask](https://github.com/dask/dask)\n",
    "    \n",
    "    - Documentation: [https://docs.dask.org](https://docs.dask.org)\n",
    "\n",
    "- Scales the existing Python ecosystem\n",
    "\n",
    "### Why Dask?\n",
    "\n",
    "- Enables parallel and larger-than-memory computations\n",
    "- Uses familiar APIs you're used to from projects like NumPy, pandas, and scikit-learn\n",
    "- Allows you to scale existing workflows with minimal code changes\n",
    "- Dask works on your laptop, but also scales out to large clusters\n",
    "- Offers great built-in diagnosic tools\n",
    "\n",
    "### Dask vs. Spark\n",
    "\n",
    "#### Reasons you might choose Spark\n",
    "\n",
    "- You prefer Scala or the SQL language\n",
    "- You have mostly JVM infrastructure and legacy systems\n",
    "- You want an established and trusted solution for business\n",
    "- You are mostly doing business analytics with some lightweight machine learning\n",
    "- You want an all-in-one solution\n",
    "\n",
    "#### Reasons you might choose Dask\n",
    "\n",
    "- You prefer Python or native code, or have large legacy code bases that you do not want to entirely rewrite\n",
    "- Your use case is complex or does not cleanly fit the Spark computing model\n",
    "- You want a lighter-weight transition from local computing to cluster computing\n",
    "- You want to interoperate with other technologies and don’t mind installing multiple packages\n",
    "\n",
    "### Components of Dask\n",
    "\n",
    "<img src=\"assets/dask-overview.png\"\n",
    "     width=\"85%\"\n",
    "     alt=\"Dask components\\\">\n",
    "     "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aac76a8f-824f-4a11-8e36-1a6981504d7a",
   "metadata": {},
   "source": [
    "# Parallelize code with `dask.delayed`\n",
    "\n",
    "In this section we parallelize simple for-loop style code with Dask and `dask.delayed`. Often, this is the only function that you will need to convert functions for use with Dask.\n",
    "\n",
    "This is a simple way to use `dask` to parallelize existing codebases or build [complex systems](https://blog.dask.org/2018/02/09/credit-models-with-dask).  This will also help us to develop an understanding for later sections.\n",
    "\n",
    "**Related Documentation**\n",
    "\n",
    "* [Delayed documentation](https://docs.dask.org/en/latest/delayed.html)\n",
    "* [Delayed screencast](https://www.youtube.com/watch?v=SHqFmynRxVU)\n",
    "* [Delayed API](https://docs.dask.org/en/latest/delayed-api.html)\n",
    "* [Delayed examples](https://examples.dask.org/delayed.html)\n",
    "* [Delayed best practices](https://docs.dask.org/en/latest/delayed-best-practices.html)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "256d2633-c71b-4576-88d2-71da0b414e7d",
   "metadata": {},
   "source": [
    "Dask has several ways of executing code in parallel. We'll use the distributed scheduler by creating a `dask.distributed.Client`. For now, this will provide us with some nice diagnostics. We'll talk about schedulers in depth later."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "de1924e7-17e6-4377-b616-75d924ce46ef",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table style=\"border: 2px solid white;\">\n",
       "<tr>\n",
       "<td style=\"vertical-align: top; border: 0px solid white\">\n",
       "<h3 style=\"text-align: left;\">Client</h3>\n",
       "<ul style=\"text-align: left; list-style: none; margin: 0; padding: 0;\">\n",
       "  <li><b>Scheduler: </b>tcp://127.0.0.1:61883</li>\n",
       "  <li><b>Dashboard: </b><a href='http://127.0.0.1:8787/status' target='_blank'>http://127.0.0.1:8787/status</a></li>\n",
       "</ul>\n",
       "</td>\n",
       "<td style=\"vertical-align: top; border: 0px solid white\">\n",
       "<h3 style=\"text-align: left;\">Cluster</h3>\n",
       "<ul style=\"text-align: left; list-style:none; margin: 0; padding: 0;\">\n",
       "  <li><b>Workers: </b>4</li>\n",
       "  <li><b>Cores: </b>16</li>\n",
       "  <li><b>Memory: </b>16.00 GiB</li>\n",
       "</ul>\n",
       "</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<Client: 'tcp://127.0.0.1:61883' processes=4 threads=16, memory=16.00 GiB>"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from dask.distributed import Client\n",
    "\n",
    "client = Client(n_workers=4)\n",
    "client"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "383281da-5a83-45b7-b969-964a3ce3858e",
   "metadata": {},
   "source": [
    "First let's make some toy functions, `inc` and `add`, that sleep for a while to simulate work. We'll then time running these functions normally."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3c8f0db8-f98a-444b-9e47-25b298dd3fb4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from time import sleep\n",
    "\n",
    "def inc(x):\n",
    "    sleep(1)\n",
    "    return x + 1\n",
    "\n",
    "def add(x, y):\n",
    "    sleep(1)\n",
    "    return x + y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f8b5fd8b-8347-4036-883b-d692af3d30f2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 285 ms, sys: 25.2 ms, total: 310 ms\n",
      "Wall time: 3.01 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# This takes three seconds to run because we call each\n",
    "# function sequentially, one after the other\n",
    "\n",
    "x = inc(1)\n",
    "y = inc(2)\n",
    "z = add(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "237bc721-3b5b-4215-8323-2376bdff95ea",
   "metadata": {},
   "source": [
    "### Parallelize with the `dask.delayed` decorator\n",
    "\n",
    "Those two increment calls *could* be called in parallel, because they are totally independent of one-another.\n",
    "\n",
    "We'll transform the `inc` and `add` functions using the `dask.delayed` function. When we call the delayed version by passing the arguments, exactly as before, the original function isn't actually called yet - which is why the cell execution finishes very quickly.\n",
    "Instead, a *delayed object* is made, which keeps track of the function to call and the arguments to pass to it.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ff175cfa-05b8-4ca8-ad9c-fa61d8d64282",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dask import delayed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "bca4f2a1-3455-40a3-a884-a7745089525c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 377 µs, sys: 106 µs, total: 483 µs\n",
      "Wall time: 443 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# This runs immediately, all it does is build a graph\n",
    "\n",
    "x = delayed(inc)(1)\n",
    "y = delayed(inc)(2)\n",
    "z = delayed(add)(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea4ae2ec-ead6-4f5a-8f28-ff77259b3f7a",
   "metadata": {},
   "source": [
    "This ran immediately, since nothing has really happened yet.\n",
    "\n",
    "To get the result, call `compute`. Notice that this runs faster than the original code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ce42f5e0-c905-452f-a754-3f3930d782f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 239 ms, sys: 19.1 ms, total: 258 ms\n",
      "Wall time: 2.03 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "# This actually runs our computation using a local thread pool\n",
    "\n",
    "z.compute()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "5489fc46-ea01-4bec-9a11-702c9f0f2e04",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Look at the task graph for `z`\n",
    "z.visualize()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c4bba1a-14d1-4df7-849d-750251605535",
   "metadata": {},
   "source": [
    "### More complex calculation\n",
    "\n",
    "To visualize how graphs can be optimally or sub-optimally built, let's look at a more complex example."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "16d7bcad-51f9-4d4a-8dda-16b8beae7208",
   "metadata": {},
   "outputs": [],
   "source": [
    "def double(x):\n",
    "    sleep(1)\n",
    "    return 2 * x\n",
    "\n",
    "def is_even(x):\n",
    "    return not x % 2\n",
    "\n",
    "data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a28fd380-c6d5-42a6-ae6a-c556c1ec5c67",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.38 s, sys: 83.6 ms, total: 1.47 s\n",
      "Wall time: 10 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "90"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "# Sequential code\n",
    "\n",
    "results = []\n",
    "for x in data:\n",
    "    if is_even(x):\n",
    "        y = double(x)\n",
    "    else:\n",
    "        y = inc(x)\n",
    "    results.append(y)\n",
    "    \n",
    "total = sum(results)\n",
    "total"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2937afd0-ed7a-4489-a40e-8b373ca88134",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 281 ms, sys: 15.5 ms, total: 296 ms\n",
      "Wall time: 1.07 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "90"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "# Parallel code -- notice I moved the `delayed` to decorators\n",
    "\n",
    "@delayed\n",
    "def inc(x):\n",
    "    sleep(1)\n",
    "    return x + 1\n",
    "\n",
    "@delayed\n",
    "def double(x):\n",
    "    sleep(1)\n",
    "    return 2 * x\n",
    "\n",
    "def is_even(x):\n",
    "    return not x % 2\n",
    "\n",
    "data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]\n",
    "\n",
    "results = []\n",
    "for x in data:\n",
    "    if is_even(x):\n",
    "        y = double(x)\n",
    "    else:\n",
    "        y = inc(x)\n",
    "    results.append(y)\n",
    "    \n",
    "total = delayed(sum)(results)\n",
    "total.compute()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "a2210af8-89e2-42d6-a5c5-78ad6c0427df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total.visualize()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbd2d207-c0d5-4d16-9073-260f56aa6389",
   "metadata": {},
   "source": [
    "### Why should you care about this?\n",
    "\n",
    "By building a specification of the calculation we want to carry out before executing anything, we can pass the specification to an *execution engine* for evaluation. In the case of Dask, this execution engine could be running on many nodes of a cluster, so you have access to the full number of CPU cores and memory across all the machines. Dask will intelligently execute your calculation with care for minimizing the amount of data held in memory, while parallelizing over the tasks that make up a graph. Notice that in the animated diagram below, where four workers are processing the (simple) graph, execution progresses vertically up the branches first, so that intermediate results can be expunged before moving onto a new branch.\n",
    "\n",
    "With `delayed` and normal pythonic looped code, very complex graphs can be built up and passed on to Dask for execution. See a nice example of [simulated complex ETL](https://blog.dask.org/2017/01/24/dask-custom) work flow.\n",
    "\n",
    "![this](assets/animation.gif)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "822a44cf-dcbd-4d9f-9b99-afe6a6c37bc9",
   "metadata": {},
   "outputs": [],
   "source": [
    "client.shutdown()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21000fcb-54a7-41c6-8a64-8e9a40074b58",
   "metadata": {},
   "source": [
    "# Bag: Parallel Lists for semi-structured data\n",
    "\n",
    "Dask-bag excels in processing data that can be represented as a sequence of arbitrary inputs. We'll refer to this as \"messy\" data, because it can contain complex nested structures, missing fields, mixtures of data types, etc. The *functional* programming style fits very nicely with standard Python iteration, such as can be found in the `itertools` module.\n",
    "\n",
    "Messy data is often encountered at the beginning of data processing pipelines when large volumes of raw data are first consumed. The initial set of data might be JSON, CSV, XML, or any other format that does not enforce strict structure and datatypes.\n",
    "For this reason, the initial data massaging and processing is often done with Python `list`s, `dict`s, and `set`s.\n",
    "\n",
    "These core data structures are optimized for general-purpose storage and processing.  Adding streaming computation with iterators/generator expressions or libraries like `itertools` or [`toolz`](https://toolz.readthedocs.io/en/latest/) let us process large volumes in a small space.  If we combine this with parallel processing then we can churn through a fair amount of data.\n",
    "\n",
    "Dask.bag is a high level Dask collection to automate common workloads of this form.  In a nutshell\n",
    "\n",
    "    dask.bag = map, filter, toolz + parallel execution\n",
    "    \n",
    "**Related Documentation**\n",
    "\n",
    "* [Bag documentation](https://docs.dask.org/en/latest/bag.html)\n",
    "* [Bag screencast](https://youtu.be/-qIiJ1XtSv0)\n",
    "* [Bag API](https://docs.dask.org/en/latest/bag-api.html)\n",
    "* [Bag examples](https://examples.dask.org/bag.html)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d45987b-077b-468d-9457-7216e3541167",
   "metadata": {},
   "source": [
    "### Create Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "eedb5ffc-7f12-4ff1-9f68-4e0a095fde1a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "distributed.client - ERROR - Failed to reconnect to scheduler after 10.00 seconds, closing client\n",
      "_GatheringFuture exception was never retrieved\n",
      "future: <_GatheringFuture finished exception=CancelledError()>\n",
      "asyncio.exceptions.CancelledError\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Created CSV acccouts in 4.01s\n",
      "Created JSON acccouts in 34.98s\n"
     ]
    }
   ],
   "source": [
    "%run prep.py -d accounts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "adffea1f-7433-49e4-a35e-d7a1dbf2a8c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dask.distributed import Client\n",
    "\n",
    "client = Client(n_workers=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3cc21dc3-75c5-4efa-be90-88f07504ef47",
   "metadata": {},
   "source": [
    "### Bag Creation\n",
    "\n",
    "You can create a `Bag` from a Python sequence, from files, from data on S3, etc.\n",
    "We demonstrate using `.take()` to show elements of the data. (Doing `.take(1)` results in a tuple with one element)\n",
    "\n",
    "Note that the data are partitioned into blocks, and there are many items per block. In the first example, the two partitions contain five elements each, and in the following two, each file is partitioned into one or more bytes blocks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d4348629-f2a7-4c15-82be-9b9e48756bcb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 2, 3)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# each element is an integer\n",
    "import dask.bag as db\n",
    "b = db.from_sequence([1, 2, 3, 4, 5, 6, 7, 8, 9, 10], npartitions=2)\n",
    "b.take(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d9ed042d-bf8c-465b-b8b9-0819be72f2d2",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'id': 1,\n",
       "  'name': 'Edith',\n",
       "  'transactions': [{'transaction-id': 107, 'amount': 7541},\n",
       "   {'transaction-id': 331, 'amount': 8108},\n",
       "   {'transaction-id': 335, 'amount': 7420},\n",
       "   {'transaction-id': 407, 'amount': 7703},\n",
       "   {'transaction-id': 489, 'amount': 7354},\n",
       "   {'transaction-id': 587, 'amount': 8334},\n",
       "   {'transaction-id': 603, 'amount': 7969},\n",
       "   {'transaction-id': 645, 'amount': 6972},\n",
       "   {'transaction-id': 707, 'amount': 7887},\n",
       "   {'transaction-id': 925, 'amount': 6884},\n",
       "   {'transaction-id': 1021, 'amount': 7287},\n",
       "   {'transaction-id': 1175, 'amount': 7960},\n",
       "   {'transaction-id': 1216, 'amount': 7495},\n",
       "   {'transaction-id': 1387, 'amount': 7433},\n",
       "   {'transaction-id': 1410, 'amount': 7182},\n",
       "   {'transaction-id': 1530, 'amount': 7991},\n",
       "   {'transaction-id': 1549, 'amount': 8005},\n",
       "   {'transaction-id': 1708, 'amount': 7974},\n",
       "   {'transaction-id': 1736, 'amount': 6878},\n",
       "   {'transaction-id': 2364, 'amount': 7667},\n",
       "   {'transaction-id': 2488, 'amount': 8358},\n",
       "   {'transaction-id': 2611, 'amount': 7908},\n",
       "   {'transaction-id': 2613, 'amount': 7614},\n",
       "   {'transaction-id': 2668, 'amount': 7519},\n",
       "   {'transaction-id': 2708, 'amount': 6751},\n",
       "   {'transaction-id': 2713, 'amount': 8414},\n",
       "   {'transaction-id': 2918, 'amount': 7838},\n",
       "   {'transaction-id': 3217, 'amount': 7903},\n",
       "   {'transaction-id': 3264, 'amount': 7578},\n",
       "   {'transaction-id': 3312, 'amount': 8449},\n",
       "   {'transaction-id': 3332, 'amount': 7247},\n",
       "   {'transaction-id': 3447, 'amount': 7114},\n",
       "   {'transaction-id': 3514, 'amount': 7424},\n",
       "   {'transaction-id': 3862, 'amount': 7895},\n",
       "   {'transaction-id': 4029, 'amount': 8032},\n",
       "   {'transaction-id': 4094, 'amount': 7671},\n",
       "   {'transaction-id': 4175, 'amount': 7517},\n",
       "   {'transaction-id': 4320, 'amount': 7505},\n",
       "   {'transaction-id': 4383, 'amount': 7706},\n",
       "   {'transaction-id': 4565, 'amount': 7834},\n",
       "   {'transaction-id': 4574, 'amount': 7241},\n",
       "   {'transaction-id': 4692, 'amount': 7343},\n",
       "   {'transaction-id': 4715, 'amount': 7009},\n",
       "   {'transaction-id': 4828, 'amount': 7996},\n",
       "   {'transaction-id': 4834, 'amount': 7676},\n",
       "   {'transaction-id': 4862, 'amount': 6898},\n",
       "   {'transaction-id': 4877, 'amount': 6913},\n",
       "   {'transaction-id': 4931, 'amount': 7305},\n",
       "   {'transaction-id': 4963, 'amount': 7658},\n",
       "   {'transaction-id': 4965, 'amount': 8428},\n",
       "   {'transaction-id': 5153, 'amount': 7965},\n",
       "   {'transaction-id': 5176, 'amount': 7683},\n",
       "   {'transaction-id': 5232, 'amount': 7485},\n",
       "   {'transaction-id': 5481, 'amount': 6154},\n",
       "   {'transaction-id': 5546, 'amount': 6693},\n",
       "   {'transaction-id': 5726, 'amount': 7995},\n",
       "   {'transaction-id': 5857, 'amount': 7735},\n",
       "   {'transaction-id': 5934, 'amount': 7730},\n",
       "   {'transaction-id': 5936, 'amount': 8188},\n",
       "   {'transaction-id': 5939, 'amount': 7120},\n",
       "   {'transaction-id': 5986, 'amount': 7743},\n",
       "   {'transaction-id': 6025, 'amount': 8305},\n",
       "   {'transaction-id': 6154, 'amount': 6619},\n",
       "   {'transaction-id': 6286, 'amount': 6619},\n",
       "   {'transaction-id': 6302, 'amount': 8054},\n",
       "   {'transaction-id': 6522, 'amount': 7515},\n",
       "   {'transaction-id': 6534, 'amount': 7224},\n",
       "   {'transaction-id': 6635, 'amount': 7628},\n",
       "   {'transaction-id': 6670, 'amount': 6447},\n",
       "   {'transaction-id': 6741, 'amount': 7873},\n",
       "   {'transaction-id': 6871, 'amount': 7516},\n",
       "   {'transaction-id': 7038, 'amount': 7904},\n",
       "   {'transaction-id': 7100, 'amount': 7514},\n",
       "   {'transaction-id': 7218, 'amount': 8362},\n",
       "   {'transaction-id': 7702, 'amount': 7368},\n",
       "   {'transaction-id': 7824, 'amount': 7709},\n",
       "   {'transaction-id': 7873, 'amount': 8198},\n",
       "   {'transaction-id': 7990, 'amount': 7013},\n",
       "   {'transaction-id': 8189, 'amount': 7694},\n",
       "   {'transaction-id': 8669, 'amount': 8199},\n",
       "   {'transaction-id': 8966, 'amount': 7726},\n",
       "   {'transaction-id': 8975, 'amount': 7579},\n",
       "   {'transaction-id': 8999, 'amount': 7519},\n",
       "   {'transaction-id': 9217, 'amount': 7890},\n",
       "   {'transaction-id': 9224, 'amount': 7056},\n",
       "   {'transaction-id': 9283, 'amount': 6977},\n",
       "   {'transaction-id': 9298, 'amount': 7174},\n",
       "   {'transaction-id': 9417, 'amount': 7126},\n",
       "   {'transaction-id': 9450, 'amount': 7890},\n",
       "   {'transaction-id': 9487, 'amount': 8375},\n",
       "   {'transaction-id': 9648, 'amount': 8375},\n",
       "   {'transaction-id': 9723, 'amount': 7182},\n",
       "   {'transaction-id': 9751, 'amount': 7077},\n",
       "   {'transaction-id': 9852, 'amount': 8055},\n",
       "   {'transaction-id': 9856, 'amount': 7938},\n",
       "   {'transaction-id': 9901, 'amount': 8055},\n",
       "   {'transaction-id': 9925, 'amount': 7691},\n",
       "   {'transaction-id': 10208, 'amount': 7215},\n",
       "   {'transaction-id': 10259, 'amount': 7628},\n",
       "   {'transaction-id': 10300, 'amount': 7620},\n",
       "   {'transaction-id': 10315, 'amount': 8160},\n",
       "   {'transaction-id': 10455, 'amount': 7461},\n",
       "   {'transaction-id': 10475, 'amount': 7788},\n",
       "   {'transaction-id': 10537, 'amount': 7539},\n",
       "   {'transaction-id': 10643, 'amount': 7107},\n",
       "   {'transaction-id': 10649, 'amount': 7112},\n",
       "   {'transaction-id': 10687, 'amount': 7688},\n",
       "   {'transaction-id': 10729, 'amount': 7974},\n",
       "   {'transaction-id': 10897, 'amount': 6782},\n",
       "   {'transaction-id': 10910, 'amount': 7442},\n",
       "   {'transaction-id': 10973, 'amount': 7330},\n",
       "   {'transaction-id': 10995, 'amount': 7899},\n",
       "   {'transaction-id': 11123, 'amount': 6924},\n",
       "   {'transaction-id': 11202, 'amount': 7708},\n",
       "   {'transaction-id': 11257, 'amount': 7140},\n",
       "   {'transaction-id': 11280, 'amount': 7719},\n",
       "   {'transaction-id': 11311, 'amount': 7935},\n",
       "   {'transaction-id': 11465, 'amount': 8560},\n",
       "   {'transaction-id': 11515, 'amount': 7618},\n",
       "   {'transaction-id': 11543, 'amount': 7932},\n",
       "   {'transaction-id': 11616, 'amount': 8720},\n",
       "   {'transaction-id': 11640, 'amount': 8020},\n",
       "   {'transaction-id': 11891, 'amount': 7921},\n",
       "   {'transaction-id': 11903, 'amount': 8082},\n",
       "   {'transaction-id': 11919, 'amount': 7256},\n",
       "   {'transaction-id': 11982, 'amount': 7927},\n",
       "   {'transaction-id': 12126, 'amount': 8017},\n",
       "   {'transaction-id': 12151, 'amount': 7279},\n",
       "   {'transaction-id': 12153, 'amount': 8124},\n",
       "   {'transaction-id': 12195, 'amount': 7216},\n",
       "   {'transaction-id': 12398, 'amount': 8015},\n",
       "   {'transaction-id': 12480, 'amount': 7452},\n",
       "   {'transaction-id': 12830, 'amount': 7367},\n",
       "   {'transaction-id': 12846, 'amount': 7135},\n",
       "   {'transaction-id': 12884, 'amount': 7729},\n",
       "   {'transaction-id': 12901, 'amount': 7444},\n",
       "   {'transaction-id': 12936, 'amount': 7225},\n",
       "   {'transaction-id': 13059, 'amount': 8648},\n",
       "   {'transaction-id': 13148, 'amount': 8149},\n",
       "   {'transaction-id': 13233, 'amount': 7637},\n",
       "   {'transaction-id': 13279, 'amount': 7721},\n",
       "   {'transaction-id': 13378, 'amount': 7469},\n",
       "   {'transaction-id': 13676, 'amount': 8065},\n",
       "   {'transaction-id': 13834, 'amount': 8403},\n",
       "   {'transaction-id': 14001, 'amount': 7278},\n",
       "   {'transaction-id': 14002, 'amount': 7768},\n",
       "   {'transaction-id': 14006, 'amount': 7990},\n",
       "   {'transaction-id': 14299, 'amount': 7970},\n",
       "   {'transaction-id': 14415, 'amount': 6928},\n",
       "   {'transaction-id': 14594, 'amount': 7726},\n",
       "   {'transaction-id': 14635, 'amount': 8024},\n",
       "   {'transaction-id': 14654, 'amount': 8003},\n",
       "   {'transaction-id': 14782, 'amount': 7768},\n",
       "   {'transaction-id': 14833, 'amount': 7928},\n",
       "   {'transaction-id': 14857, 'amount': 7827},\n",
       "   {'transaction-id': 14859, 'amount': 8677},\n",
       "   {'transaction-id': 14871, 'amount': 7406},\n",
       "   {'transaction-id': 15051, 'amount': 7661},\n",
       "   {'transaction-id': 15148, 'amount': 7691},\n",
       "   {'transaction-id': 15349, 'amount': 7394},\n",
       "   {'transaction-id': 15642, 'amount': 7654},\n",
       "   {'transaction-id': 15703, 'amount': 6621},\n",
       "   {'transaction-id': 15785, 'amount': 7646},\n",
       "   {'transaction-id': 15933, 'amount': 7776},\n",
       "   {'transaction-id': 16006, 'amount': 8555},\n",
       "   {'transaction-id': 16047, 'amount': 7546},\n",
       "   {'transaction-id': 16325, 'amount': 6875},\n",
       "   {'transaction-id': 16485, 'amount': 8003},\n",
       "   {'transaction-id': 16574, 'amount': 7372},\n",
       "   {'transaction-id': 16622, 'amount': 8186},\n",
       "   {'transaction-id': 16712, 'amount': 7759},\n",
       "   {'transaction-id': 16762, 'amount': 7527},\n",
       "   {'transaction-id': 16872, 'amount': 7589},\n",
       "   {'transaction-id': 17025, 'amount': 7338},\n",
       "   {'transaction-id': 17234, 'amount': 7582},\n",
       "   {'transaction-id': 17283, 'amount': 7711},\n",
       "   {'transaction-id': 17436, 'amount': 7954},\n",
       "   {'transaction-id': 17575, 'amount': 7796},\n",
       "   {'transaction-id': 17717, 'amount': 7724},\n",
       "   {'transaction-id': 17803, 'amount': 8548},\n",
       "   {'transaction-id': 17907, 'amount': 8551},\n",
       "   {'transaction-id': 18026, 'amount': 7532},\n",
       "   {'transaction-id': 18027, 'amount': 6814},\n",
       "   {'transaction-id': 18338, 'amount': 6796},\n",
       "   {'transaction-id': 18347, 'amount': 7340},\n",
       "   {'transaction-id': 18356, 'amount': 7843},\n",
       "   {'transaction-id': 18636, 'amount': 8041},\n",
       "   {'transaction-id': 18655, 'amount': 7431},\n",
       "   {'transaction-id': 18707, 'amount': 7385},\n",
       "   {'transaction-id': 19107, 'amount': 7733},\n",
       "   {'transaction-id': 19121, 'amount': 7712},\n",
       "   {'transaction-id': 19197, 'amount': 7286},\n",
       "   {'transaction-id': 19203, 'amount': 7282},\n",
       "   {'transaction-id': 19248, 'amount': 6603},\n",
       "   {'transaction-id': 19463, 'amount': 7265},\n",
       "   {'transaction-id': 19724, 'amount': 7804},\n",
       "   {'transaction-id': 19832, 'amount': 7454},\n",
       "   {'transaction-id': 19890, 'amount': 8259},\n",
       "   {'transaction-id': 19895, 'amount': 7007},\n",
       "   {'transaction-id': 19950, 'amount': 6447},\n",
       "   {'transaction-id': 20044, 'amount': 7972},\n",
       "   {'transaction-id': 20121, 'amount': 7081},\n",
       "   {'transaction-id': 20390, 'amount': 7370},\n",
       "   {'transaction-id': 20464, 'amount': 8313},\n",
       "   {'transaction-id': 20809, 'amount': 6875},\n",
       "   {'transaction-id': 20813, 'amount': 7942},\n",
       "   {'transaction-id': 20905, 'amount': 7742},\n",
       "   {'transaction-id': 20983, 'amount': 8829},\n",
       "   {'transaction-id': 21022, 'amount': 7802},\n",
       "   {'transaction-id': 21150, 'amount': 7672},\n",
       "   {'transaction-id': 21486, 'amount': 7403},\n",
       "   {'transaction-id': 21502, 'amount': 8679},\n",
       "   {'transaction-id': 21582, 'amount': 7067},\n",
       "   {'transaction-id': 21592, 'amount': 7425},\n",
       "   {'transaction-id': 21761, 'amount': 7229},\n",
       "   {'transaction-id': 21877, 'amount': 6886},\n",
       "   {'transaction-id': 22073, 'amount': 7609},\n",
       "   {'transaction-id': 22143, 'amount': 7589},\n",
       "   {'transaction-id': 22219, 'amount': 7264},\n",
       "   {'transaction-id': 22337, 'amount': 7109},\n",
       "   {'transaction-id': 22630, 'amount': 7682},\n",
       "   {'transaction-id': 22655, 'amount': 8161},\n",
       "   {'transaction-id': 22714, 'amount': 7920},\n",
       "   {'transaction-id': 22915, 'amount': 8165},\n",
       "   {'transaction-id': 23298, 'amount': 8027},\n",
       "   {'transaction-id': 23422, 'amount': 8932},\n",
       "   {'transaction-id': 23470, 'amount': 7685},\n",
       "   {'transaction-id': 23473, 'amount': 7836},\n",
       "   {'transaction-id': 23507, 'amount': 6698},\n",
       "   {'transaction-id': 23526, 'amount': 6962},\n",
       "   {'transaction-id': 23579, 'amount': 7535},\n",
       "   {'transaction-id': 23628, 'amount': 7426},\n",
       "   {'transaction-id': 23751, 'amount': 7765},\n",
       "   {'transaction-id': 23792, 'amount': 8138},\n",
       "   {'transaction-id': 23857, 'amount': 7557},\n",
       "   {'transaction-id': 23898, 'amount': 7082},\n",
       "   {'transaction-id': 23961, 'amount': 7236},\n",
       "   {'transaction-id': 23981, 'amount': 7384},\n",
       "   {'transaction-id': 24226, 'amount': 8342},\n",
       "   {'transaction-id': 24244, 'amount': 7929},\n",
       "   {'transaction-id': 24245, 'amount': 7644},\n",
       "   {'transaction-id': 24340, 'amount': 7803},\n",
       "   {'transaction-id': 24433, 'amount': 8284},\n",
       "   {'transaction-id': 24518, 'amount': 8444},\n",
       "   {'transaction-id': 24584, 'amount': 8001},\n",
       "   {'transaction-id': 24723, 'amount': 8153},\n",
       "   {'transaction-id': 24962, 'amount': 7482},\n",
       "   {'transaction-id': 25004, 'amount': 7282},\n",
       "   {'transaction-id': 25066, 'amount': 7826},\n",
       "   {'transaction-id': 25067, 'amount': 7783},\n",
       "   {'transaction-id': 25306, 'amount': 7905},\n",
       "   {'transaction-id': 25723, 'amount': 8510},\n",
       "   {'transaction-id': 25735, 'amount': 8037},\n",
       "   {'transaction-id': 26014, 'amount': 7755},\n",
       "   {'transaction-id': 26041, 'amount': 7680},\n",
       "   {'transaction-id': 26272, 'amount': 7457},\n",
       "   {'transaction-id': 26324, 'amount': 8395},\n",
       "   {'transaction-id': 26565, 'amount': 8255},\n",
       "   {'transaction-id': 26869, 'amount': 7427},\n",
       "   {'transaction-id': 26897, 'amount': 8298},\n",
       "   {'transaction-id': 27032, 'amount': 8683},\n",
       "   {'transaction-id': 27057, 'amount': 8111},\n",
       "   {'transaction-id': 27255, 'amount': 7707},\n",
       "   {'transaction-id': 27577, 'amount': 7474},\n",
       "   {'transaction-id': 27723, 'amount': 7619},\n",
       "   {'transaction-id': 27899, 'amount': 7344},\n",
       "   {'transaction-id': 27919, 'amount': 7475},\n",
       "   {'transaction-id': 27956, 'amount': 7844},\n",
       "   {'transaction-id': 28053, 'amount': 8301},\n",
       "   {'transaction-id': 28159, 'amount': 8067},\n",
       "   {'transaction-id': 28207, 'amount': 7929},\n",
       "   {'transaction-id': 28278, 'amount': 8363},\n",
       "   {'transaction-id': 28305, 'amount': 7783},\n",
       "   {'transaction-id': 28307, 'amount': 7818},\n",
       "   {'transaction-id': 28368, 'amount': 7379},\n",
       "   {'transaction-id': 28453, 'amount': 7292},\n",
       "   {'transaction-id': 28798, 'amount': 8618},\n",
       "   {'transaction-id': 29051, 'amount': 8233},\n",
       "   {'transaction-id': 29114, 'amount': 7914},\n",
       "   {'transaction-id': 29204, 'amount': 7737},\n",
       "   {'transaction-id': 29309, 'amount': 7665},\n",
       "   {'transaction-id': 29379, 'amount': 8214},\n",
       "   {'transaction-id': 29395, 'amount': 6955},\n",
       "   {'transaction-id': 29478, 'amount': 7522},\n",
       "   {'transaction-id': 29480, 'amount': 7479},\n",
       "   {'transaction-id': 30068, 'amount': 7425},\n",
       "   {'transaction-id': 30130, 'amount': 7179},\n",
       "   {'transaction-id': 30156, 'amount': 7407},\n",
       "   {'transaction-id': 30177, 'amount': 7503},\n",
       "   {'transaction-id': 30199, 'amount': 7212},\n",
       "   {'transaction-id': 30277, 'amount': 8077},\n",
       "   {'transaction-id': 30281, 'amount': 7786},\n",
       "   {'transaction-id': 30333, 'amount': 7214},\n",
       "   {'transaction-id': 30655, 'amount': 7796},\n",
       "   {'transaction-id': 30688, 'amount': 7922},\n",
       "   {'transaction-id': 30698, 'amount': 7336},\n",
       "   {'transaction-id': 30784, 'amount': 6809},\n",
       "   {'transaction-id': 30893, 'amount': 7574},\n",
       "   {'transaction-id': 30990, 'amount': 7530},\n",
       "   {'transaction-id': 31028, 'amount': 7497},\n",
       "   {'transaction-id': 31206, 'amount': 7559},\n",
       "   {'transaction-id': 31419, 'amount': 8164},\n",
       "   {'transaction-id': 31723, 'amount': 7173},\n",
       "   {'transaction-id': 31927, 'amount': 7736},\n",
       "   {'transaction-id': 32072, 'amount': 7715},\n",
       "   {'transaction-id': 32077, 'amount': 7278},\n",
       "   {'transaction-id': 32108, 'amount': 6978},\n",
       "   {'transaction-id': 32143, 'amount': 7552},\n",
       "   {'transaction-id': 32666, 'amount': 7401},\n",
       "   {'transaction-id': 32697, 'amount': 8077},\n",
       "   {'transaction-id': 32766, 'amount': 7788},\n",
       "   {'transaction-id': 32977, 'amount': 7839},\n",
       "   {'transaction-id': 32997, 'amount': 8169},\n",
       "   {'transaction-id': 33110, 'amount': 8225},\n",
       "   {'transaction-id': 33161, 'amount': 7775},\n",
       "   {'transaction-id': 33164, 'amount': 7000},\n",
       "   {'transaction-id': 33318, 'amount': 7702},\n",
       "   {'transaction-id': 33509, 'amount': 8317},\n",
       "   {'transaction-id': 33518, 'amount': 7537},\n",
       "   {'transaction-id': 33535, 'amount': 7748},\n",
       "   {'transaction-id': 33544, 'amount': 7491},\n",
       "   {'transaction-id': 33716, 'amount': 7499},\n",
       "   {'transaction-id': 33781, 'amount': 7517},\n",
       "   {'transaction-id': 33891, 'amount': 7170},\n",
       "   {'transaction-id': 34102, 'amount': 7131},\n",
       "   {'transaction-id': 34357, 'amount': 8587},\n",
       "   {'transaction-id': 34670, 'amount': 7348},\n",
       "   {'transaction-id': 34711, 'amount': 7105},\n",
       "   {'transaction-id': 34729, 'amount': 8513},\n",
       "   {'transaction-id': 34861, 'amount': 6971},\n",
       "   {'transaction-id': 34982, 'amount': 8134},\n",
       "   {'transaction-id': 35072, 'amount': 7945},\n",
       "   {'transaction-id': 35348, 'amount': 7465},\n",
       "   {'transaction-id': 35393, 'amount': 7315},\n",
       "   {'transaction-id': 35454, 'amount': 7520},\n",
       "   {'transaction-id': 35531, 'amount': 7411},\n",
       "   {'transaction-id': 35532, 'amount': 7691},\n",
       "   {'transaction-id': 35715, 'amount': 7618},\n",
       "   {'transaction-id': 35770, 'amount': 8477},\n",
       "   {'transaction-id': 35779, 'amount': 7950},\n",
       "   {'transaction-id': 35846, 'amount': 7903},\n",
       "   {'transaction-id': 36110, 'amount': 6917},\n",
       "   {'transaction-id': 36250, 'amount': 7446},\n",
       "   {'transaction-id': 36256, 'amount': 7557},\n",
       "   {'transaction-id': 36503, 'amount': 7384},\n",
       "   {'transaction-id': 36594, 'amount': 7958},\n",
       "   {'transaction-id': 36696, 'amount': 7801},\n",
       "   {'transaction-id': 36857, 'amount': 7930},\n",
       "   {'transaction-id': 37118, 'amount': 7674},\n",
       "   {'transaction-id': 37139, 'amount': 7404},\n",
       "   {'transaction-id': 37158, 'amount': 7899},\n",
       "   {'transaction-id': 37477, 'amount': 6979},\n",
       "   {'transaction-id': 37566, 'amount': 7849},\n",
       "   {'transaction-id': 37604, 'amount': 7499},\n",
       "   {'transaction-id': 37688, 'amount': 7196},\n",
       "   {'transaction-id': 37701, 'amount': 7848},\n",
       "   {'transaction-id': 38242, 'amount': 7118},\n",
       "   {'transaction-id': 38249, 'amount': 7899},\n",
       "   {'transaction-id': 38306, 'amount': 7097},\n",
       "   {'transaction-id': 38313, 'amount': 7845},\n",
       "   {'transaction-id': 38434, 'amount': 7021},\n",
       "   {'transaction-id': 38465, 'amount': 7282},\n",
       "   {'transaction-id': 38662, 'amount': 8232},\n",
       "   {'transaction-id': 38689, 'amount': 6927},\n",
       "   {'transaction-id': 38734, 'amount': 7450},\n",
       "   {'transaction-id': 38867, 'amount': 8062},\n",
       "   {'transaction-id': 38883, 'amount': 8188},\n",
       "   {'transaction-id': 38965, 'amount': 7604},\n",
       "   {'transaction-id': 39057, 'amount': 7602},\n",
       "   {'transaction-id': 39265, 'amount': 8089},\n",
       "   {'transaction-id': 39305, 'amount': 7119},\n",
       "   {'transaction-id': 39330, 'amount': 7831},\n",
       "   {'transaction-id': 39403, 'amount': 7677},\n",
       "   {'transaction-id': 39410, 'amount': 8397},\n",
       "   {'transaction-id': 39491, 'amount': 7250},\n",
       "   {'transaction-id': 39664, 'amount': 7031},\n",
       "   {'transaction-id': 39698, 'amount': 7564},\n",
       "   {'transaction-id': 39762, 'amount': 7021},\n",
       "   {'transaction-id': 39799, 'amount': 7915},\n",
       "   {'transaction-id': 39830, 'amount': 7714},\n",
       "   {'transaction-id': 39983, 'amount': 7864},\n",
       "   {'transaction-id': 40003, 'amount': 7846},\n",
       "   {'transaction-id': 40060, 'amount': 7247},\n",
       "   {'transaction-id': 40164, 'amount': 7044},\n",
       "   {'transaction-id': 40271, 'amount': 8328},\n",
       "   {'transaction-id': 40408, 'amount': 7325},\n",
       "   {'transaction-id': 40549, 'amount': 7880},\n",
       "   {'transaction-id': 40555, 'amount': 7960},\n",
       "   {'transaction-id': 40572, 'amount': 7978},\n",
       "   {'transaction-id': 40689, 'amount': 7549},\n",
       "   {'transaction-id': 40950, 'amount': 7526},\n",
       "   {'transaction-id': 41050, 'amount': 6818},\n",
       "   {'transaction-id': 41125, 'amount': 7909},\n",
       "   {'transaction-id': 41199, 'amount': 7953},\n",
       "   {'transaction-id': 41230, 'amount': 8543},\n",
       "   {'transaction-id': 41404, 'amount': 7041},\n",
       "   {'transaction-id': 41425, 'amount': 7588},\n",
       "   {'transaction-id': 41522, 'amount': 8019},\n",
       "   {'transaction-id': 41535, 'amount': 7120},\n",
       "   {'transaction-id': 41658, 'amount': 7442},\n",
       "   {'transaction-id': 41745, 'amount': 7660},\n",
       "   {'transaction-id': 41754, 'amount': 7427},\n",
       "   {'transaction-id': 41813, 'amount': 7387},\n",
       "   {'transaction-id': 41831, 'amount': 6546},\n",
       "   {'transaction-id': 41857, 'amount': 7288},\n",
       "   {'transaction-id': 41859, 'amount': 7618},\n",
       "   {'transaction-id': 41883, 'amount': 7719},\n",
       "   {'transaction-id': 42041, 'amount': 7797},\n",
       "   {'transaction-id': 42378, 'amount': 7812},\n",
       "   {'transaction-id': 42430, 'amount': 7877},\n",
       "   {'transaction-id': 42432, 'amount': 7698},\n",
       "   {'transaction-id': 42463, 'amount': 7469},\n",
       "   {'transaction-id': 42489, 'amount': 8120},\n",
       "   {'transaction-id': 42526, 'amount': 7216},\n",
       "   {'transaction-id': 42529, 'amount': 7334},\n",
       "   {'transaction-id': 42546, 'amount': 8214},\n",
       "   {'transaction-id': 42654, 'amount': 8037},\n",
       "   {'transaction-id': 42850, 'amount': 7525},\n",
       "   {'transaction-id': 42880, 'amount': 7773},\n",
       "   {'transaction-id': 42943, 'amount': 7525},\n",
       "   {'transaction-id': 43144, 'amount': 7495},\n",
       "   {'transaction-id': 43184, 'amount': 7509},\n",
       "   {'transaction-id': 43306, 'amount': 7373},\n",
       "   {'transaction-id': 43391, 'amount': 7412},\n",
       "   {'transaction-id': 43454, 'amount': 8382},\n",
       "   {'transaction-id': 43494, 'amount': 7198},\n",
       "   {'transaction-id': 43507, 'amount': 7798},\n",
       "   {'transaction-id': 43827, 'amount': 7381},\n",
       "   {'transaction-id': 43834, 'amount': 7519},\n",
       "   {'transaction-id': 43839, 'amount': 8297},\n",
       "   {'transaction-id': 43895, 'amount': 7298},\n",
       "   {'transaction-id': 44069, 'amount': 7102},\n",
       "   {'transaction-id': 44072, 'amount': 7142},\n",
       "   {'transaction-id': 44076, 'amount': 7380},\n",
       "   {'transaction-id': 44165, 'amount': 8606},\n",
       "   {'transaction-id': 44228, 'amount': 8393},\n",
       "   {'transaction-id': 44329, 'amount': 7992},\n",
       "   {'transaction-id': 44495, 'amount': 7502},\n",
       "   {'transaction-id': 44583, 'amount': 7298},\n",
       "   {'transaction-id': 44944, 'amount': 7658},\n",
       "   {'transaction-id': 45006, 'amount': 7662},\n",
       "   {'transaction-id': 45094, 'amount': 8055},\n",
       "   {'transaction-id': 45242, 'amount': 7438},\n",
       "   {'transaction-id': 45437, 'amount': 7642},\n",
       "   {'transaction-id': 45603, 'amount': 7648},\n",
       "   {'transaction-id': 45649, 'amount': 7045},\n",
       "   {'transaction-id': 45735, 'amount': 7654},\n",
       "   {'transaction-id': 45771, 'amount': 7532},\n",
       "   {'transaction-id': 45857, 'amount': 7894},\n",
       "   {'transaction-id': 45966, 'amount': 8045},\n",
       "   {'transaction-id': 46009, 'amount': 6467},\n",
       "   {'transaction-id': 46041, 'amount': 7627},\n",
       "   {'transaction-id': 46183, 'amount': 7399},\n",
       "   {'transaction-id': 46420, 'amount': 7416},\n",
       "   {'transaction-id': 46482, 'amount': 7189},\n",
       "   {'transaction-id': 46628, 'amount': 7410},\n",
       "   {'transaction-id': 46690, 'amount': 6867},\n",
       "   {'transaction-id': 46827, 'amount': 8208},\n",
       "   {'transaction-id': 46938, 'amount': 7438},\n",
       "   {'transaction-id': 46987, 'amount': 7481},\n",
       "   {'transaction-id': 47107, 'amount': 7620},\n",
       "   {'transaction-id': 47521, 'amount': 7500},\n",
       "   {'transaction-id': 48018, 'amount': 7948},\n",
       "   {'transaction-id': 48043, 'amount': 7663},\n",
       "   {'transaction-id': 48206, 'amount': 7470},\n",
       "   {'transaction-id': 48349, 'amount': 7932},\n",
       "   {'transaction-id': 48545, 'amount': 7195},\n",
       "   {'transaction-id': 48641, 'amount': 7392},\n",
       "   {'transaction-id': 48778, 'amount': 7743},\n",
       "   {'transaction-id': 48906, 'amount': 8118},\n",
       "   {'transaction-id': 48949, 'amount': 7849},\n",
       "   {'transaction-id': 49538, 'amount': 6741},\n",
       "   {'transaction-id': 49685, 'amount': 6858},\n",
       "   {'transaction-id': 49884, 'amount': 7468},\n",
       "   {'transaction-id': 49887, 'amount': 7819},\n",
       "   {'transaction-id': 49935, 'amount': 8384},\n",
       "   {'transaction-id': 50059, 'amount': 7792},\n",
       "   {'transaction-id': 50079, 'amount': 7872},\n",
       "   {'transaction-id': 50199, 'amount': 8178},\n",
       "   {'transaction-id': 50290, 'amount': 8043},\n",
       "   {'transaction-id': 50400, 'amount': 7282},\n",
       "   {'transaction-id': 50452, 'amount': 7449},\n",
       "   {'transaction-id': 50457, 'amount': 7212},\n",
       "   {'transaction-id': 50653, 'amount': 7329},\n",
       "   {'transaction-id': 50703, 'amount': 8308},\n",
       "   {'transaction-id': 50927, 'amount': 8733},\n",
       "   {'transaction-id': 50964, 'amount': 7563},\n",
       "   {'transaction-id': 51174, 'amount': 7625},\n",
       "   {'transaction-id': 51214, 'amount': 7850},\n",
       "   {'transaction-id': 51226, 'amount': 7138},\n",
       "   {'transaction-id': 51230, 'amount': 7538},\n",
       "   {'transaction-id': 51285, 'amount': 8003},\n",
       "   {'transaction-id': 51597, 'amount': 7060},\n",
       "   {'transaction-id': 51651, 'amount': 7025},\n",
       "   {'transaction-id': 51718, 'amount': 7371},\n",
       "   {'transaction-id': 51769, 'amount': 7328},\n",
       "   {'transaction-id': 51842, 'amount': 7510},\n",
       "   {'transaction-id': 51865, 'amount': 6808},\n",
       "   {'transaction-id': 51920, 'amount': 7661},\n",
       "   {'transaction-id': 51949, 'amount': 7639},\n",
       "   {'transaction-id': 52039, 'amount': 8248},\n",
       "   {'transaction-id': 52067, 'amount': 7612},\n",
       "   {'transaction-id': 52127, 'amount': 8154},\n",
       "   {'transaction-id': 52325, 'amount': 8361},\n",
       "   {'transaction-id': 52488, 'amount': 7974},\n",
       "   {'transaction-id': 52532, 'amount': 8163},\n",
       "   {'transaction-id': 52626, 'amount': 8180},\n",
       "   {'transaction-id': 52813, 'amount': 7138},\n",
       "   {'transaction-id': 52817, 'amount': 7195},\n",
       "   {'transaction-id': 52893, 'amount': 7646},\n",
       "   {'transaction-id': 52912, 'amount': 8524},\n",
       "   {'transaction-id': 53213, 'amount': 7150},\n",
       "   {'transaction-id': 53217, 'amount': 7606},\n",
       "   {'transaction-id': 53301, 'amount': 7685},\n",
       "   {'transaction-id': 53408, 'amount': 7414},\n",
       "   {'transaction-id': 53534, 'amount': 6952},\n",
       "   {'transaction-id': 53568, 'amount': 8358},\n",
       "   {'transaction-id': 53622, 'amount': 6118},\n",
       "   {'transaction-id': 53762, 'amount': 7940},\n",
       "   {'transaction-id': 53846, 'amount': 7414},\n",
       "   {'transaction-id': 53887, 'amount': 8054},\n",
       "   {'transaction-id': 53985, 'amount': 8092},\n",
       "   {'transaction-id': 54148, 'amount': 7366},\n",
       "   {'transaction-id': 54172, 'amount': 7728},\n",
       "   {'transaction-id': 54320, 'amount': 7603},\n",
       "   {'transaction-id': 54445, 'amount': 7864},\n",
       "   {'transaction-id': 54561, 'amount': 7263},\n",
       "   {'transaction-id': 54684, 'amount': 7704},\n",
       "   {'transaction-id': 54695, 'amount': 8527},\n",
       "   {'transaction-id': 54732, 'amount': 7900},\n",
       "   {'transaction-id': 54896, 'amount': 7695},\n",
       "   {'transaction-id': 54899, 'amount': 7424},\n",
       "   {'transaction-id': 55020, 'amount': 7254},\n",
       "   {'transaction-id': 55126, 'amount': 7106},\n",
       "   {'transaction-id': 55356, 'amount': 7446},\n",
       "   {'transaction-id': 55440, 'amount': 7146},\n",
       "   {'transaction-id': 55627, 'amount': 7718},\n",
       "   {'transaction-id': 55658, 'amount': 8487},\n",
       "   {'transaction-id': 55758, 'amount': 7825},\n",
       "   {'transaction-id': 55896, 'amount': 7120},\n",
       "   {'transaction-id': 55966, 'amount': 7162},\n",
       "   {'transaction-id': 56080, 'amount': 7151},\n",
       "   {'transaction-id': 56156, 'amount': 7521},\n",
       "   {'transaction-id': 56194, 'amount': 7009},\n",
       "   {'transaction-id': 56209, 'amount': 7693},\n",
       "   {'transaction-id': 56382, 'amount': 7204},\n",
       "   {'transaction-id': 56394, 'amount': 7318},\n",
       "   {'transaction-id': 56484, 'amount': 7959},\n",
       "   {'transaction-id': 56697, 'amount': 7007},\n",
       "   {'transaction-id': 56722, 'amount': 7742},\n",
       "   {'transaction-id': 56855, 'amount': 7607},\n",
       "   {'transaction-id': 57275, 'amount': 7754},\n",
       "   {'transaction-id': 57310, 'amount': 7512},\n",
       "   {'transaction-id': 57379, 'amount': 7862},\n",
       "   {'transaction-id': 57419, 'amount': 7027},\n",
       "   {'transaction-id': 57498, 'amount': 7385},\n",
       "   {'transaction-id': 57608, 'amount': 7306},\n",
       "   {'transaction-id': 57685, 'amount': 6745},\n",
       "   {'transaction-id': 57739, 'amount': 7328},\n",
       "   {'transaction-id': 57778, 'amount': 7938},\n",
       "   {'transaction-id': 57887, 'amount': 7741},\n",
       "   {'transaction-id': 57920, 'amount': 7418},\n",
       "   {'transaction-id': 58071, 'amount': 7665},\n",
       "   {'transaction-id': 58115, 'amount': 7758},\n",
       "   {'transaction-id': 58559, 'amount': 7424},\n",
       "   {'transaction-id': 58694, 'amount': 7572},\n",
       "   {'transaction-id': 58724, 'amount': 8008},\n",
       "   {'transaction-id': 58783, 'amount': 8157},\n",
       "   {'transaction-id': 58803, 'amount': 7320},\n",
       "   {'transaction-id': 58829, 'amount': 8064},\n",
       "   {'transaction-id': 58881, 'amount': 7234},\n",
       "   {'transaction-id': 58885, 'amount': 7643},\n",
       "   {'transaction-id': 58925, 'amount': 7218},\n",
       "   {'transaction-id': 58926, 'amount': 7669},\n",
       "   {'transaction-id': 58948, 'amount': 7694},\n",
       "   {'transaction-id': 58982, 'amount': 6825},\n",
       "   {'transaction-id': 59412, 'amount': 8258},\n",
       "   {'transaction-id': 59549, 'amount': 7603},\n",
       "   {'transaction-id': 59598, 'amount': 7461},\n",
       "   {'transaction-id': 59600, 'amount': 7350},\n",
       "   {'transaction-id': 59618, 'amount': 7664},\n",
       "   {'transaction-id': 59920, 'amount': 7915},\n",
       "   {'transaction-id': 59924, 'amount': 8251},\n",
       "   {'transaction-id': 60119, 'amount': 7430},\n",
       "   {'transaction-id': 60194, 'amount': 7213},\n",
       "   {'transaction-id': 60224, 'amount': 7470},\n",
       "   {'transaction-id': 60452, 'amount': 7196},\n",
       "   {'transaction-id': 60587, 'amount': 8027},\n",
       "   {'transaction-id': 60592, 'amount': 7163},\n",
       "   {'transaction-id': 60617, 'amount': 7497},\n",
       "   {'transaction-id': 60790, 'amount': 7899},\n",
       "   {'transaction-id': 60810, 'amount': 7652},\n",
       "   {'transaction-id': 60943, 'amount': 7294},\n",
       "   {'transaction-id': 61117, 'amount': 7194},\n",
       "   {'transaction-id': 61156, 'amount': 6556},\n",
       "   {'transaction-id': 61235, 'amount': 7585},\n",
       "   {'transaction-id': 61366, 'amount': 7730},\n",
       "   {'transaction-id': 61467, 'amount': 8091},\n",
       "   {'transaction-id': 61468, 'amount': 8175},\n",
       "   {'transaction-id': 61504, 'amount': 7412},\n",
       "   {'transaction-id': 61861, 'amount': 7737},\n",
       "   {'transaction-id': 62036, 'amount': 8263},\n",
       "   {'transaction-id': 62071, 'amount': 8194},\n",
       "   {'transaction-id': 62159, 'amount': 7265},\n",
       "   {'transaction-id': 62250, 'amount': 8125},\n",
       "   {'transaction-id': 62555, 'amount': 8311},\n",
       "   {'transaction-id': 62722, 'amount': 7448},\n",
       "   {'transaction-id': 62772, 'amount': 7779},\n",
       "   {'transaction-id': 62864, 'amount': 7393},\n",
       "   {'transaction-id': 63137, 'amount': 8290},\n",
       "   {'transaction-id': 63193, 'amount': 7903},\n",
       "   {'transaction-id': 63230, 'amount': 7215},\n",
       "   {'transaction-id': 63252, 'amount': 6935},\n",
       "   {'transaction-id': 63263, 'amount': 8224},\n",
       "   {'transaction-id': 63302, 'amount': 7536},\n",
       "   {'transaction-id': 63559, 'amount': 7756},\n",
       "   {'transaction-id': 63596, 'amount': 6961},\n",
       "   {'transaction-id': 63644, 'amount': 7800},\n",
       "   {'transaction-id': 63719, 'amount': 7179},\n",
       "   {'transaction-id': 63801, 'amount': 7102},\n",
       "   {'transaction-id': 64021, 'amount': 7901},\n",
       "   {'transaction-id': 64129, 'amount': 6684},\n",
       "   {'transaction-id': 64193, 'amount': 7120},\n",
       "   {'transaction-id': 64199, 'amount': 6939},\n",
       "   {'transaction-id': 64222, 'amount': 7717},\n",
       "   {'transaction-id': 64292, 'amount': 7555},\n",
       "   {'transaction-id': 64458, 'amount': 7881},\n",
       "   {'transaction-id': 64575, 'amount': 6841},\n",
       "   {'transaction-id': 64632, 'amount': 7338},\n",
       "   {'transaction-id': 64701, 'amount': 7845},\n",
       "   {'transaction-id': 64940, 'amount': 7174},\n",
       "   {'transaction-id': 65005, 'amount': 8084},\n",
       "   {'transaction-id': 65058, 'amount': 7996},\n",
       "   {'transaction-id': 65143, 'amount': 6599},\n",
       "   {'transaction-id': 65200, 'amount': 7722},\n",
       "   {'transaction-id': 65296, 'amount': 7694},\n",
       "   {'transaction-id': 65487, 'amount': 7771},\n",
       "   {'transaction-id': 65532, 'amount': 7513},\n",
       "   {'transaction-id': 65599, 'amount': 7298},\n",
       "   {'transaction-id': 65622, 'amount': 7283},\n",
       "   {'transaction-id': 65624, 'amount': 7881},\n",
       "   {'transaction-id': 65632, 'amount': 7940},\n",
       "   {'transaction-id': 65777, 'amount': 7577},\n",
       "   {'transaction-id': 65847, 'amount': 7260},\n",
       "   {'transaction-id': 66154, 'amount': 7425},\n",
       "   {'transaction-id': 66195, 'amount': 6621},\n",
       "   {'transaction-id': 66352, 'amount': 7778},\n",
       "   {'transaction-id': 66427, 'amount': 7720},\n",
       "   {'transaction-id': 66436, 'amount': 7539},\n",
       "   {'transaction-id': 66460, 'amount': 7363},\n",
       "   {'transaction-id': 66494, 'amount': 7990},\n",
       "   {'transaction-id': 66641, 'amount': 7654},\n",
       "   {'transaction-id': 66703, 'amount': 8050},\n",
       "   {'transaction-id': 66862, 'amount': 7799},\n",
       "   {'transaction-id': 66879, 'amount': 7159},\n",
       "   {'transaction-id': 66967, 'amount': 6503},\n",
       "   {'transaction-id': 67238, 'amount': 7401},\n",
       "   {'transaction-id': 67266, 'amount': 7110},\n",
       "   {'transaction-id': 67381, 'amount': 7410},\n",
       "   {'transaction-id': 67417, 'amount': 8387},\n",
       "   {'transaction-id': 67470, 'amount': 8007},\n",
       "   {'transaction-id': 67518, 'amount': 7818},\n",
       "   {'transaction-id': 67710, 'amount': 6593},\n",
       "   {'transaction-id': 67727, 'amount': 8013},\n",
       "   {'transaction-id': 67919, 'amount': 7852},\n",
       "   {'transaction-id': 67938, 'amount': 7111},\n",
       "   {'transaction-id': 67964, 'amount': 7138},\n",
       "   {'transaction-id': 68475, 'amount': 8184},\n",
       "   {'transaction-id': 68573, 'amount': 6596},\n",
       "   {'transaction-id': 68885, 'amount': 7723},\n",
       "   {'transaction-id': 69188, 'amount': 7953},\n",
       "   {'transaction-id': 69217, 'amount': 7799},\n",
       "   {'transaction-id': 69464, 'amount': 7569},\n",
       "   {'transaction-id': 69496, 'amount': 7712},\n",
       "   {'transaction-id': 69627, 'amount': 7419},\n",
       "   {'transaction-id': 69765, 'amount': 7344},\n",
       "   {'transaction-id': 69785, 'amount': 7414},\n",
       "   {'transaction-id': 69792, 'amount': 7254},\n",
       "   {'transaction-id': 69928, 'amount': 6873},\n",
       "   {'transaction-id': 69959, 'amount': 7241},\n",
       "   {'transaction-id': 70017, 'amount': 6764},\n",
       "   {'transaction-id': 70047, 'amount': 7225},\n",
       "   {'transaction-id': 70085, 'amount': 8204},\n",
       "   {'transaction-id': 70129, 'amount': 7623},\n",
       "   {'transaction-id': 70183, 'amount': 7047},\n",
       "   {'transaction-id': 70405, 'amount': 7773},\n",
       "   {'transaction-id': 70469, 'amount': 7732},\n",
       "   {'transaction-id': 70544, 'amount': 7907},\n",
       "   {'transaction-id': 71043, 'amount': 6835},\n",
       "   {'transaction-id': 71494, 'amount': 7015},\n",
       "   {'transaction-id': 71602, 'amount': 8453},\n",
       "   {'transaction-id': 71662, 'amount': 7211},\n",
       "   {'transaction-id': 71752, 'amount': 7763},\n",
       "   {'transaction-id': 71758, 'amount': 8524},\n",
       "   {'transaction-id': 71841, 'amount': 7119},\n",
       "   {'transaction-id': 72064, 'amount': 7446},\n",
       "   {'transaction-id': 72199, 'amount': 7206},\n",
       "   {'transaction-id': 72265, 'amount': 8034},\n",
       "   {'transaction-id': 72268, 'amount': 7519},\n",
       "   {'transaction-id': 72368, 'amount': 7916},\n",
       "   {'transaction-id': 72431, 'amount': 8109},\n",
       "   {'transaction-id': 72577, 'amount': 6724},\n",
       "   {'transaction-id': 72946, 'amount': 7447},\n",
       "   {'transaction-id': 73133, 'amount': 8419},\n",
       "   {'transaction-id': 73262, 'amount': 8035},\n",
       "   {'transaction-id': 73529, 'amount': 7380},\n",
       "   {'transaction-id': 73568, 'amount': 8101},\n",
       "   {'transaction-id': 73734, 'amount': 8927},\n",
       "   {'transaction-id': 73831, 'amount': 7370},\n",
       "   {'transaction-id': 73918, 'amount': 8030},\n",
       "   {'transaction-id': 74003, 'amount': 7894},\n",
       "   {'transaction-id': 74037, 'amount': 7426},\n",
       "   {'transaction-id': 74132, 'amount': 6593},\n",
       "   {'transaction-id': 74197, 'amount': 6992},\n",
       "   {'transaction-id': 74242, 'amount': 7236},\n",
       "   {'transaction-id': 74370, 'amount': 7481},\n",
       "   {'transaction-id': 74450, 'amount': 7752},\n",
       "   {'transaction-id': 74549, 'amount': 7292},\n",
       "   {'transaction-id': 74627, 'amount': 7000},\n",
       "   {'transaction-id': 74729, 'amount': 7870},\n",
       "   {'transaction-id': 74750, 'amount': 8409},\n",
       "   {'transaction-id': 74762, 'amount': 8236},\n",
       "   {'transaction-id': 74782, 'amount': 7801},\n",
       "   {'transaction-id': 74871, 'amount': 7548},\n",
       "   {'transaction-id': 74873, 'amount': 7808},\n",
       "   {'transaction-id': 74897, 'amount': 8132},\n",
       "   {'transaction-id': 75021, 'amount': 7763},\n",
       "   {'transaction-id': 75057, 'amount': 8624},\n",
       "   {'transaction-id': 75183, 'amount': 6584},\n",
       "   {'transaction-id': 75194, 'amount': 7301},\n",
       "   {'transaction-id': 75246, 'amount': 7384},\n",
       "   {'transaction-id': 75262, 'amount': 6976},\n",
       "   {'transaction-id': 75429, 'amount': 7364},\n",
       "   {'transaction-id': 75631, 'amount': 7071},\n",
       "   {'transaction-id': 75730, 'amount': 7533},\n",
       "   {'transaction-id': 76031, 'amount': 6804},\n",
       "   {'transaction-id': 76135, 'amount': 8066},\n",
       "   {'transaction-id': 76213, 'amount': 8116},\n",
       "   {'transaction-id': 76262, 'amount': 7631},\n",
       "   {'transaction-id': 76310, 'amount': 6824},\n",
       "   {'transaction-id': 76317, 'amount': 7961},\n",
       "   {'transaction-id': 76428, 'amount': 8376},\n",
       "   {'transaction-id': 76472, 'amount': 8082},\n",
       "   {'transaction-id': 76531, 'amount': 7389},\n",
       "   {'transaction-id': 76579, 'amount': 8458},\n",
       "   {'transaction-id': 76587, 'amount': 7662},\n",
       "   {'transaction-id': 76701, 'amount': 7669},\n",
       "   {'transaction-id': 76826, 'amount': 6299},\n",
       "   {'transaction-id': 76874, 'amount': 7858},\n",
       "   {'transaction-id': 76929, 'amount': 7828},\n",
       "   {'transaction-id': 77207, 'amount': 7399},\n",
       "   {'transaction-id': 77226, 'amount': 7718},\n",
       "   {'transaction-id': 77379, 'amount': 7594},\n",
       "   {'transaction-id': 77461, 'amount': 6500},\n",
       "   {'transaction-id': 77525, 'amount': 6970},\n",
       "   {'transaction-id': 77576, 'amount': 8033},\n",
       "   {'transaction-id': 77679, 'amount': 7259},\n",
       "   {'transaction-id': 77823, 'amount': 7570},\n",
       "   {'transaction-id': 77835, 'amount': 7563},\n",
       "   {'transaction-id': 77842, 'amount': 7473},\n",
       "   {'transaction-id': 78120, 'amount': 8487},\n",
       "   {'transaction-id': 78127, 'amount': 7214},\n",
       "   {'transaction-id': 78296, 'amount': 7543},\n",
       "   {'transaction-id': 78389, 'amount': 7077},\n",
       "   {'transaction-id': 78445, 'amount': 7662},\n",
       "   {'transaction-id': 78509, 'amount': 7309},\n",
       "   {'transaction-id': 78569, 'amount': 7953},\n",
       "   {'transaction-id': 78585, 'amount': 7579},\n",
       "   {'transaction-id': 78596, 'amount': 8348},\n",
       "   {'transaction-id': 78650, 'amount': 6923},\n",
       "   {'transaction-id': 78653, 'amount': 6772},\n",
       "   {'transaction-id': 78660, 'amount': 8161},\n",
       "   {'transaction-id': 78724, 'amount': 8591},\n",
       "   {'transaction-id': 78988, 'amount': 6716},\n",
       "   {'transaction-id': 79076, 'amount': 8122},\n",
       "   {'transaction-id': 79169, 'amount': 8133},\n",
       "   {'transaction-id': 79279, 'amount': 7819},\n",
       "   {'transaction-id': 79445, 'amount': 7839},\n",
       "   {'transaction-id': 79511, 'amount': 8102},\n",
       "   {'transaction-id': 79621, 'amount': 7790},\n",
       "   {'transaction-id': 79675, 'amount': 7152},\n",
       "   {'transaction-id': 80124, 'amount': 7880},\n",
       "   {'transaction-id': 80375, 'amount': 7265},\n",
       "   {'transaction-id': 80483, 'amount': 8085},\n",
       "   {'transaction-id': 80594, 'amount': 7665},\n",
       "   {'transaction-id': 80652, 'amount': 7555},\n",
       "   {'transaction-id': 80713, 'amount': 7526},\n",
       "   {'transaction-id': 80906, 'amount': 7598},\n",
       "   {'transaction-id': 80965, 'amount': 6935},\n",
       "   {'transaction-id': 81011, 'amount': 7285},\n",
       "   {'transaction-id': 81149, 'amount': 6942},\n",
       "   {'transaction-id': 81204, 'amount': 8358},\n",
       "   {'transaction-id': 81635, 'amount': 8387},\n",
       "   {'transaction-id': 81716, 'amount': 7744},\n",
       "   {'transaction-id': 81788, 'amount': 7730},\n",
       "   {'transaction-id': 81859, 'amount': 7693},\n",
       "   {'transaction-id': 82003, 'amount': 8088},\n",
       "   {'transaction-id': 82080, 'amount': 7514},\n",
       "   {'transaction-id': 82197, 'amount': 7794},\n",
       "   {'transaction-id': 82350, 'amount': 7549},\n",
       "   {'transaction-id': 82367, 'amount': 7324},\n",
       "   {'transaction-id': 82518, 'amount': 8166},\n",
       "   {'transaction-id': 82812, 'amount': 7813},\n",
       "   {'transaction-id': 82986, 'amount': 7577},\n",
       "   {'transaction-id': 83128, 'amount': 8226},\n",
       "   {'transaction-id': 83258, 'amount': 7565},\n",
       "   {'transaction-id': 83457, 'amount': 7291},\n",
       "   {'transaction-id': 83468, 'amount': 6651},\n",
       "   {'transaction-id': 83470, 'amount': 6713},\n",
       "   {'transaction-id': 83501, 'amount': 7622},\n",
       "   {'transaction-id': 83769, 'amount': 7261},\n",
       "   {'transaction-id': 83887, 'amount': 7640},\n",
       "   {'transaction-id': 83920, 'amount': 8498},\n",
       "   {'transaction-id': 83947, 'amount': 7230},\n",
       "   {'transaction-id': 84051, 'amount': 7500},\n",
       "   {'transaction-id': 84128, 'amount': 7503},\n",
       "   {'transaction-id': 84283, 'amount': 6942},\n",
       "   {'transaction-id': 84423, 'amount': 8593},\n",
       "   {'transaction-id': 84469, 'amount': 6791},\n",
       "   {'transaction-id': 84609, 'amount': 8354},\n",
       "   {'transaction-id': 84620, 'amount': 7888},\n",
       "   {'transaction-id': 84681, 'amount': 7100},\n",
       "   {'transaction-id': 84818, 'amount': 7541},\n",
       "   {'transaction-id': 84827, 'amount': 7641},\n",
       "   {'transaction-id': 84931, 'amount': 7789},\n",
       "   {'transaction-id': 85003, 'amount': 7832},\n",
       "   {'transaction-id': 85073, 'amount': 8186},\n",
       "   {'transaction-id': 85180, 'amount': 7596},\n",
       "   {'transaction-id': 85242, 'amount': 7394},\n",
       "   {'transaction-id': 85357, 'amount': 7532},\n",
       "   {'transaction-id': 85363, 'amount': 7784},\n",
       "   {'transaction-id': 85412, 'amount': 7938},\n",
       "   {'transaction-id': 85423, 'amount': 7787},\n",
       "   {'transaction-id': 85731, 'amount': 7627},\n",
       "   {'transaction-id': 85797, 'amount': 7089},\n",
       "   {'transaction-id': 85802, 'amount': 8230},\n",
       "   {'transaction-id': 85940, 'amount': 7561},\n",
       "   {'transaction-id': 86042, 'amount': 7173},\n",
       "   {'transaction-id': 86078, 'amount': 7062},\n",
       "   {'transaction-id': 86227, 'amount': 7546},\n",
       "   {'transaction-id': 86305, 'amount': 6971},\n",
       "   {'transaction-id': 86351, 'amount': 8360},\n",
       "   {'transaction-id': 86357, 'amount': 7777},\n",
       "   {'transaction-id': 86717, 'amount': 7244},\n",
       "   {'transaction-id': 86928, 'amount': 7833},\n",
       "   {'transaction-id': 86998, 'amount': 7808},\n",
       "   {'transaction-id': 87049, 'amount': 7464},\n",
       "   {'transaction-id': 87324, 'amount': 7810},\n",
       "   {'transaction-id': 87478, 'amount': 8301},\n",
       "   {'transaction-id': 87542, 'amount': 7915},\n",
       "   {'transaction-id': 87545, 'amount': 7216},\n",
       "   {'transaction-id': 87643, 'amount': 7520},\n",
       "   {'transaction-id': 87884, 'amount': 7729},\n",
       "   {'transaction-id': 88033, 'amount': 6690},\n",
       "   {'transaction-id': 88047, 'amount': 7073},\n",
       "   {'transaction-id': 88142, 'amount': 7591},\n",
       "   {'transaction-id': 88181, 'amount': 7785},\n",
       "   {'transaction-id': 88183, 'amount': 7708},\n",
       "   {'transaction-id': 88659, 'amount': 7303},\n",
       "   {'transaction-id': 88791, 'amount': 6988},\n",
       "   {'transaction-id': 88817, 'amount': 7773},\n",
       "   {'transaction-id': 88885, 'amount': 7276},\n",
       "   {'transaction-id': 88959, 'amount': 7955},\n",
       "   {'transaction-id': 88977, 'amount': 7238},\n",
       "   {'transaction-id': 89105, 'amount': 7370},\n",
       "   {'transaction-id': 89203, 'amount': 7722},\n",
       "   {'transaction-id': 89214, 'amount': 7536},\n",
       "   {'transaction-id': 89219, 'amount': 6797},\n",
       "   {'transaction-id': 89262, 'amount': 7557},\n",
       "   {'transaction-id': 89553, 'amount': 7833},\n",
       "   {'transaction-id': 89702, 'amount': 6889},\n",
       "   {'transaction-id': 89712, 'amount': 8086},\n",
       "   {'transaction-id': 89973, 'amount': 8317},\n",
       "   {'transaction-id': 90104, 'amount': 7590},\n",
       "   {'transaction-id': 90110, 'amount': 8099},\n",
       "   {'transaction-id': 90331, 'amount': 7845},\n",
       "   {'transaction-id': 90362, 'amount': 7189},\n",
       "   {'transaction-id': 90377, 'amount': 7847},\n",
       "   {'transaction-id': 90390, 'amount': 7752},\n",
       "   {'transaction-id': 90652, 'amount': 7078},\n",
       "   {'transaction-id': 90746, 'amount': 7689},\n",
       "   {'transaction-id': 90889, 'amount': 7637},\n",
       "   {'transaction-id': 90945, 'amount': 7438},\n",
       "   {'transaction-id': 91176, 'amount': 7704},\n",
       "   {'transaction-id': 91684, 'amount': 7338},\n",
       "   {'transaction-id': 92002, 'amount': 8310},\n",
       "   {'transaction-id': 92124, 'amount': 6631},\n",
       "   {'transaction-id': 92152, 'amount': 8171},\n",
       "   {'transaction-id': 92316, 'amount': 7320},\n",
       "   {'transaction-id': 92340, 'amount': 8192},\n",
       "   {'transaction-id': 92356, 'amount': 7703},\n",
       "   {'transaction-id': 92438, 'amount': 7743},\n",
       "   {'transaction-id': 92451, 'amount': 7877},\n",
       "   {'transaction-id': 92682, 'amount': 7892},\n",
       "   {'transaction-id': 92924, 'amount': 8407},\n",
       "   {'transaction-id': 93202, 'amount': 7721},\n",
       "   {'transaction-id': 93225, 'amount': 7508},\n",
       "   {'transaction-id': 93299, 'amount': 8457},\n",
       "   {'transaction-id': 93365, 'amount': 7526},\n",
       "   {'transaction-id': 93512, 'amount': 7576},\n",
       "   {'transaction-id': 93583, 'amount': 7920},\n",
       "   {'transaction-id': 93589, 'amount': 7897},\n",
       "   {'transaction-id': 93602, 'amount': 7619},\n",
       "   {'transaction-id': 93669, 'amount': 8666},\n",
       "   {'transaction-id': 93730, 'amount': 8253},\n",
       "   {'transaction-id': 93770, 'amount': 6937},\n",
       "   {'transaction-id': 93892, 'amount': 7610},\n",
       "   {'transaction-id': 94029, 'amount': 8004},\n",
       "   {'transaction-id': 94084, 'amount': 8019},\n",
       "   {'transaction-id': 94094, 'amount': 7740},\n",
       "   {'transaction-id': 94172, 'amount': 7673},\n",
       "   {'transaction-id': 94461, 'amount': 7401},\n",
       "   {'transaction-id': 94515, 'amount': 8051},\n",
       "   {'transaction-id': 94567, 'amount': 7360},\n",
       "   {'transaction-id': 94700, 'amount': 8115},\n",
       "   {'transaction-id': 94824, 'amount': 8013},\n",
       "   {'transaction-id': 94851, 'amount': 7963},\n",
       "   {'transaction-id': 94901, 'amount': 7037},\n",
       "   {'transaction-id': 95065, 'amount': 7600},\n",
       "   {'transaction-id': 95162, 'amount': 7235},\n",
       "   {'transaction-id': 95356, 'amount': 7520},\n",
       "   {'transaction-id': 95428, 'amount': 7238},\n",
       "   {'transaction-id': 95500, 'amount': 6846},\n",
       "   {'transaction-id': 95790, 'amount': 8091},\n",
       "   {'transaction-id': 95793, 'amount': 7984},\n",
       "   {'transaction-id': 95853, 'amount': 7013},\n",
       "   {'transaction-id': 95979, 'amount': 7270},\n",
       "   {'transaction-id': 96001, 'amount': 8155},\n",
       "   {'transaction-id': 96071, 'amount': 6769},\n",
       "   {'transaction-id': 96091, 'amount': 7469},\n",
       "   {'transaction-id': 96195, 'amount': 7213},\n",
       "   {'transaction-id': 96224, 'amount': 8181},\n",
       "   {'transaction-id': 96443, 'amount': 7116},\n",
       "   {'transaction-id': 96725, 'amount': 8578},\n",
       "   {'transaction-id': 96770, 'amount': 7097},\n",
       "   {'transaction-id': 96856, 'amount': 7480},\n",
       "   {'transaction-id': 97191, 'amount': 7811},\n",
       "   {'transaction-id': 97371, 'amount': 7290},\n",
       "   {'transaction-id': 97434, 'amount': 7809},\n",
       "   {'transaction-id': 97442, 'amount': 8034},\n",
       "   {'transaction-id': 97510, 'amount': 7460},\n",
       "   {'transaction-id': 97560, 'amount': 6787},\n",
       "   {'transaction-id': 97678, 'amount': 7083},\n",
       "   {'transaction-id': 97890, 'amount': 7391},\n",
       "   {'transaction-id': 97939, 'amount': 7386},\n",
       "   {'transaction-id': 97993, 'amount': 7120},\n",
       "   {'transaction-id': 98008, 'amount': 7550},\n",
       "   {'transaction-id': 98032, 'amount': 7383},\n",
       "   {'transaction-id': 98226, 'amount': 7842},\n",
       "   {'transaction-id': 98233, 'amount': 7569},\n",
       "   {'transaction-id': 98255, 'amount': 7385},\n",
       "   {'transaction-id': 98740, 'amount': 7666},\n",
       "   {'transaction-id': 98776, 'amount': 7454},\n",
       "   {'transaction-id': 98809, 'amount': 6843},\n",
       "   {'transaction-id': 98826, 'amount': 7844},\n",
       "   {'transaction-id': 98872, 'amount': 8021},\n",
       "   {'transaction-id': 98955, 'amount': 7427},\n",
       "   {'transaction-id': 99069, 'amount': 7682},\n",
       "   {'transaction-id': 99149, 'amount': 8233},\n",
       "   {'transaction-id': 99281, 'amount': 8212},\n",
       "   {'transaction-id': 99416, 'amount': 7636},\n",
       "   {'transaction-id': 99476, 'amount': 7830},\n",
       "   {'transaction-id': 99700, 'amount': 8139},\n",
       "   {'transaction-id': 99799, 'amount': 7776},\n",
       "   {'transaction-id': 99988, 'amount': 8134}]},)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# each element is a text file, where each line is a JSON object\n",
    "# note that the compression is handled automatically\n",
    "import os\n",
    "import json\n",
    "\n",
    "lines = db.read_text(os.path.join('data', 'accounts.*.json.gz'))\n",
    "js = lines.map(json.loads)\n",
    "# take: inspect first few elements\n",
    "js.take(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd43e807-85b6-4c68-86a2-3f4e06c93614",
   "metadata": {},
   "source": [
    "We will use the fake dataset we've created of gzipped JSON data in your data directory.  This is like the example used in the `DataFrame` example we will see later, except that it has bundled up all of the entries for each individual `id` into a single record.  This is similar to data that you might collect off of a document store database or a web API.\n",
    "\n",
    "Each line is a JSON encoded dictionary with the following keys\n",
    "\n",
    "*  id: Unique identifier of the customer\n",
    "*  name: Name of the customer\n",
    "*  transactions: List of `transaction-id`, `amount` pairs, one for each transaction for the customer in that file"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0fdc4aa2-e369-407f-863f-3ed4f1287cf4",
   "metadata": {},
   "source": [
    "Once we parse our JSON data into proper Python objects (`dict`s, `list`s, etc.) we can perform more interesting queries by creating small Python functions to run on our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "92659d62-fa14-4654-8933-4364d496e88c",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'id': 37,\n",
       "  'name': 'Alice',\n",
       "  'transactions': [{'transaction-id': 472, 'amount': 174},\n",
       "   {'transaction-id': 1537, 'amount': 165},\n",
       "   {'transaction-id': 3208, 'amount': 172},\n",
       "   {'transaction-id': 3361, 'amount': 169},\n",
       "   {'transaction-id': 4212, 'amount': 174},\n",
       "   {'transaction-id': 4286, 'amount': 176},\n",
       "   {'transaction-id': 5901, 'amount': 178},\n",
       "   {'transaction-id': 7399, 'amount': 173},\n",
       "   {'transaction-id': 7714, 'amount': 163},\n",
       "   {'transaction-id': 9613, 'amount': 175},\n",
       "   {'transaction-id': 9758, 'amount': 169},\n",
       "   {'transaction-id': 10522, 'amount': 174},\n",
       "   {'transaction-id': 10983, 'amount': 175},\n",
       "   {'transaction-id': 11747, 'amount': 174},\n",
       "   {'transaction-id': 12210, 'amount': 170},\n",
       "   {'transaction-id': 12476, 'amount': 169},\n",
       "   {'transaction-id': 12542, 'amount': 179},\n",
       "   {'transaction-id': 13122, 'amount': 174},\n",
       "   {'transaction-id': 13867, 'amount': 175},\n",
       "   {'transaction-id': 14309, 'amount': 175},\n",
       "   {'transaction-id': 14981, 'amount': 167},\n",
       "   {'transaction-id': 15922, 'amount': 174},\n",
       "   {'transaction-id': 17648, 'amount': 176},\n",
       "   {'transaction-id': 17708, 'amount': 163},\n",
       "   {'transaction-id': 17861, 'amount': 162},\n",
       "   {'transaction-id': 18834, 'amount': 172},\n",
       "   {'transaction-id': 20338, 'amount': 170},\n",
       "   {'transaction-id': 20514, 'amount': 168},\n",
       "   {'transaction-id': 20559, 'amount': 170},\n",
       "   {'transaction-id': 20615, 'amount': 180},\n",
       "   {'transaction-id': 21382, 'amount': 178},\n",
       "   {'transaction-id': 21407, 'amount': 176},\n",
       "   {'transaction-id': 23068, 'amount': 160},\n",
       "   {'transaction-id': 24102, 'amount': 153},\n",
       "   {'transaction-id': 24349, 'amount': 171},\n",
       "   {'transaction-id': 24452, 'amount': 166},\n",
       "   {'transaction-id': 26280, 'amount': 174},\n",
       "   {'transaction-id': 27108, 'amount': 167},\n",
       "   {'transaction-id': 29223, 'amount': 167},\n",
       "   {'transaction-id': 29880, 'amount': 168},\n",
       "   {'transaction-id': 30076, 'amount': 161},\n",
       "   {'transaction-id': 30224, 'amount': 167},\n",
       "   {'transaction-id': 30744, 'amount': 169},\n",
       "   {'transaction-id': 31520, 'amount': 171},\n",
       "   {'transaction-id': 32323, 'amount': 168},\n",
       "   {'transaction-id': 32808, 'amount': 162},\n",
       "   {'transaction-id': 33004, 'amount': 167},\n",
       "   {'transaction-id': 33098, 'amount': 183},\n",
       "   {'transaction-id': 33375, 'amount': 189},\n",
       "   {'transaction-id': 34595, 'amount': 184},\n",
       "   {'transaction-id': 34620, 'amount': 162},\n",
       "   {'transaction-id': 35815, 'amount': 176},\n",
       "   {'transaction-id': 36184, 'amount': 160},\n",
       "   {'transaction-id': 39632, 'amount': 175},\n",
       "   {'transaction-id': 41195, 'amount': 166},\n",
       "   {'transaction-id': 42243, 'amount': 167},\n",
       "   {'transaction-id': 42804, 'amount': 169},\n",
       "   {'transaction-id': 43973, 'amount': 182},\n",
       "   {'transaction-id': 43977, 'amount': 150},\n",
       "   {'transaction-id': 44967, 'amount': 169},\n",
       "   {'transaction-id': 46193, 'amount': 168},\n",
       "   {'transaction-id': 46491, 'amount': 165},\n",
       "   {'transaction-id': 46570, 'amount': 159},\n",
       "   {'transaction-id': 47370, 'amount': 180},\n",
       "   {'transaction-id': 48375, 'amount': 176},\n",
       "   {'transaction-id': 49396, 'amount': 173},\n",
       "   {'transaction-id': 50392, 'amount': 167},\n",
       "   {'transaction-id': 50446, 'amount': 161},\n",
       "   {'transaction-id': 51503, 'amount': 176},\n",
       "   {'transaction-id': 52862, 'amount': 176},\n",
       "   {'transaction-id': 53984, 'amount': 167},\n",
       "   {'transaction-id': 54044, 'amount': 176},\n",
       "   {'transaction-id': 54644, 'amount': 163},\n",
       "   {'transaction-id': 54965, 'amount': 184},\n",
       "   {'transaction-id': 55056, 'amount': 159},\n",
       "   {'transaction-id': 56234, 'amount': 164},\n",
       "   {'transaction-id': 56405, 'amount': 157},\n",
       "   {'transaction-id': 56543, 'amount': 177},\n",
       "   {'transaction-id': 57118, 'amount': 162},\n",
       "   {'transaction-id': 58310, 'amount': 176},\n",
       "   {'transaction-id': 59869, 'amount': 155},\n",
       "   {'transaction-id': 60348, 'amount': 168},\n",
       "   {'transaction-id': 61209, 'amount': 175},\n",
       "   {'transaction-id': 64235, 'amount': 163},\n",
       "   {'transaction-id': 64449, 'amount': 157},\n",
       "   {'transaction-id': 64812, 'amount': 176},\n",
       "   {'transaction-id': 65180, 'amount': 170},\n",
       "   {'transaction-id': 65463, 'amount': 169},\n",
       "   {'transaction-id': 65817, 'amount': 170},\n",
       "   {'transaction-id': 66294, 'amount': 163},\n",
       "   {'transaction-id': 66592, 'amount': 172},\n",
       "   {'transaction-id': 66813, 'amount': 177},\n",
       "   {'transaction-id': 67342, 'amount': 165},\n",
       "   {'transaction-id': 67396, 'amount': 169},\n",
       "   {'transaction-id': 68399, 'amount': 177},\n",
       "   {'transaction-id': 68869, 'amount': 176},\n",
       "   {'transaction-id': 68954, 'amount': 182},\n",
       "   {'transaction-id': 70551, 'amount': 165},\n",
       "   {'transaction-id': 71604, 'amount': 172},\n",
       "   {'transaction-id': 71642, 'amount': 161},\n",
       "   {'transaction-id': 73008, 'amount': 180},\n",
       "   {'transaction-id': 75152, 'amount': 163},\n",
       "   {'transaction-id': 76329, 'amount': 175},\n",
       "   {'transaction-id': 78411, 'amount': 174},\n",
       "   {'transaction-id': 78648, 'amount': 167},\n",
       "   {'transaction-id': 78649, 'amount': 162},\n",
       "   {'transaction-id': 79141, 'amount': 176},\n",
       "   {'transaction-id': 79602, 'amount': 159},\n",
       "   {'transaction-id': 81040, 'amount': 164},\n",
       "   {'transaction-id': 83144, 'amount': 165},\n",
       "   {'transaction-id': 83686, 'amount': 182},\n",
       "   {'transaction-id': 84308, 'amount': 172},\n",
       "   {'transaction-id': 84885, 'amount': 172},\n",
       "   {'transaction-id': 85978, 'amount': 169},\n",
       "   {'transaction-id': 86072, 'amount': 168},\n",
       "   {'transaction-id': 86365, 'amount': 172},\n",
       "   {'transaction-id': 88089, 'amount': 170},\n",
       "   {'transaction-id': 88101, 'amount': 176},\n",
       "   {'transaction-id': 89540, 'amount': 175},\n",
       "   {'transaction-id': 89541, 'amount': 172},\n",
       "   {'transaction-id': 90596, 'amount': 171},\n",
       "   {'transaction-id': 90735, 'amount': 177},\n",
       "   {'transaction-id': 91017, 'amount': 162},\n",
       "   {'transaction-id': 91598, 'amount': 176},\n",
       "   {'transaction-id': 92868, 'amount': 178},\n",
       "   {'transaction-id': 93641, 'amount': 164},\n",
       "   {'transaction-id': 93817, 'amount': 169},\n",
       "   {'transaction-id': 94392, 'amount': 162},\n",
       "   {'transaction-id': 94806, 'amount': 176},\n",
       "   {'transaction-id': 96840, 'amount': 167},\n",
       "   {'transaction-id': 97512, 'amount': 168},\n",
       "   {'transaction-id': 97919, 'amount': 168},\n",
       "   {'transaction-id': 98202, 'amount': 172},\n",
       "   {'transaction-id': 99952, 'amount': 165}]},)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# filter: keep only some elements of the sequence\n",
    "js.filter(lambda record: record['name'] == 'Alice').take(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "6ce2e3e1-d2a8-490b-8547-af5306fe0bc4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'name': 'Alice', 'count': 134},\n",
       " {'name': 'Alice', 'count': 319},\n",
       " {'name': 'Alice', 'count': 220},\n",
       " {'name': 'Alice', 'count': 18},\n",
       " {'name': 'Alice', 'count': 136})"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def count_transactions(d):\n",
    "    return {'name': d['name'], 'count': len(d['transactions'])}\n",
    "\n",
    "# map: apply a function to each element\n",
    "(js.filter(lambda record: record['name'] == 'Alice')\n",
    "   .map(count_transactions)\n",
    "   .take(5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "11f74049-c8de-4e8b-9b12-ed5dcc53f80d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(134, 319, 220, 18, 136)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pluck: select a field, as from a dictionary, element[field]\n",
    "(js.filter(lambda record: record['name'] == 'Alice')\n",
    "   .map(count_transactions)\n",
    "   .pluck('count')\n",
    "   .take(5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "5515e689-0fb9-41d2-8dc0-a94328734e13",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "204.026"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Average number of transactions for all of the Alice entries\n",
    "(js.filter(lambda record: record['name'] == 'Alice')\n",
    "   .map(count_transactions)\n",
    "   .pluck('count')\n",
    "   .mean()\n",
    "   .compute())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d627183e-4e71-4632-96ec-c2c2a09e8627",
   "metadata": {},
   "outputs": [],
   "source": [
    "client.shutdown()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94d42dae-e477-4fb3-a1de-dc0ec8efb3ff",
   "metadata": {},
   "source": [
    "# Arrays: analogous to Numpy Arrays\n",
    "\n",
    "<img src=\"assets/dask-array.png\" width=\"50%\">\n",
    "\n",
    "Dask array provides a parallel, larger-than-memory, n-dimensional array using blocked algorithms. Simply put: distributed Numpy.\n",
    "\n",
    "*  **Parallel**: Uses all of the cores on your computer\n",
    "*  **Larger-than-memory**:  Lets you work on datasets that are larger than your available memory by breaking up your array into many small pieces, operating on those pieces in an order that minimizes the memory footprint of your computation, and effectively streaming data from disk.\n",
    "*  **Blocked Algorithms**:  Perform large computations by performing many smaller computations\n",
    "\n",
    "Here, we'll build some understanding by implementing some blocked algorithms from scratch.\n",
    "We'll then use Dask Array to analyze large datasets, in parallel, using a familiar NumPy-like API.\n",
    "\n",
    "**Related Documentation**\n",
    "\n",
    "* [Array documentation](https://docs.dask.org/en/latest/array.html)\n",
    "* [Array screencast](https://youtu.be/9h_61hXCDuI)\n",
    "* [Array API](https://docs.dask.org/en/latest/array-api.html)\n",
    "* [Array examples](https://examples.dask.org/array.html)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "586a2c6a-1b42-4178-89d8-d3cc6ba4ce31",
   "metadata": {},
   "source": [
    "### Create Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "35a90d12-b422-497a-be8c-af52bcbb690e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "distributed.client - ERROR - Failed to reconnect to scheduler after 10.00 seconds, closing client\n",
      "_GatheringFuture exception was never retrieved\n",
      "future: <_GatheringFuture finished exception=CancelledError()>\n",
      "asyncio.exceptions.CancelledError\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Created random data for array exercise in 19.81s\n"
     ]
    }
   ],
   "source": [
    "%run prep.py -d random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "159384fe-861e-47c0-bd06-8c69aeea9092",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dask.distributed import Client\n",
    "\n",
    "client = Client(n_workers=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "70209b0c-65c4-4ebe-b921-3253aaf36e55",
   "metadata": {},
   "source": [
    "### Blocked Algorithms\n",
    "\n",
    "A *blocked algorithm* executes on a large dataset by breaking it up into many small blocks.\n",
    "\n",
    "For example, consider taking the sum of a billion numbers.  We might instead break up the array into 1,000 chunks, each of size 1,000,000, take the sum of each chunk, and then take the sum of the intermediate sums.\n",
    "\n",
    "We achieve the intended result (one sum on one billion numbers) by performing many smaller results (one thousand sums on one million numbers each, followed by another sum of a thousand numbers.)\n",
    "\n",
    "We do exactly this with Python and NumPy in the following example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "48068ce6-0623-4cbd-acfd-1fb699a6f69c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data with h5py\n",
    "# this creates a pointer to the data, but does not actually load\n",
    "import h5py\n",
    "import os\n",
    "f = h5py.File(os.path.join('data', 'random.hdf5'), mode='r')\n",
    "dset = f['/x']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "490e8bd7-e5e5-4f15-ad68-f8fb0ad343b8",
   "metadata": {},
   "source": [
    "#### Compute Sum Using Blocked Algorithm\n",
    "\n",
    "Before using dask, let's consider the concept of blocked algorithms. We can compute the sum of a large number of elements by loading them chunk-by-chunk, and keeping a running total.\n",
    "\n",
    "Here we compute the sum of this large array on disk by \n",
    "\n",
    "1.  Computing the sum of each 1,000,000 sized chunk of the array\n",
    "2.  Computing the sum of the 1,000 intermediate sums\n",
    "\n",
    "Note that this is a sequential process in the notebook kernel, both the loading and summing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "63cd3105-4ca8-4580-ba4d-131fb943b084",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 883 ms, sys: 1 s, total: 1.89 s\n",
      "Wall time: 1.95 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1000010204.0625"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# Compute sum of large array, one million numbers at a time\n",
    "sums = []\n",
    "for i in range(0, 1_000_000_000, 1_000_000):\n",
    "    chunk = dset[i: i + 1_000_000]  # pull out numpy array\n",
    "    sums.append(chunk.sum())\n",
    "\n",
    "total = sum(sums)\n",
    "total"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "381d53ab-1ccf-4aa0-92fe-f936d35d611c",
   "metadata": {},
   "source": [
    "#### Compute Mean Using Blocked Algorithm\n",
    "\n",
    "Now that we've seen the simple example above, try doing a slightly more complicated problem. Compute the mean of the array, assuming for a moment that we don't happen to already know how many elements are in the data.  You can do this by changing the code above with the following alterations:\n",
    "\n",
    "1.  Compute the sum of each block\n",
    "2.  Compute the length of each block\n",
    "3.  Compute the sum of the 1,000 intermediate sums and the sum of the 1,000 intermediate lengths and divide one by the other\n",
    "\n",
    "This approach is overkill for our case but does nicely generalize if we don't know the size of the array or individual blocks beforehand."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "624756f5-b659-4a73-bceb-d926a9e85b80",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 683 ms, sys: 535 ms, total: 1.22 s\n",
      "Wall time: 1.19 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1.0000102040625"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "sums = []\n",
    "lengths = []\n",
    "for i in range(0, 1_000_000_000, 1_000_000):\n",
    "    chunk = dset[i: i + 1_000_000]  # pull out numpy array\n",
    "    sums.append(chunk.sum())\n",
    "    lengths.append(len(chunk))\n",
    "\n",
    "total = sum(sums)\n",
    "length = sum(lengths)\n",
    "total / length"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fbba0393-3709-452a-937c-0cb7e5827ced",
   "metadata": {},
   "source": [
    "### `dask.array` contains these algorithms\n",
    "\n",
    "Dask.array is a NumPy-like library that does these kinds of tricks to operate on large datasets that don't fit into memory.  It extends beyond the linear problems discussed above to full N-Dimensional algorithms and a decent subset of the NumPy interface.\n",
    "\n",
    "You can create a `dask.array` `Array` object with the `da.from_array` function.  This function accepts\n",
    "\n",
    "1.  `data`: Any object that supports NumPy slicing, like `dset`\n",
    "2.  `chunks`: A chunk size to tell us how to block up our array, like `(1_000_000,)`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "466020a0-5d79-411f-948d-86f2d2db59fd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<tr>\n",
       "<td>\n",
       "<table>\n",
       "  <thead>\n",
       "    <tr><td> </td><th> Array </th><th> Chunk </th></tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr><th> Bytes </th><td> 3.73 GiB </td> <td> 3.81 MiB </td></tr>\n",
       "    <tr><th> Shape </th><td> (1000000000,) </td> <td> (1000000,) </td></tr>\n",
       "    <tr><th> Count </th><td> 1001 Tasks </td><td> 1000 Chunks </td></tr>\n",
       "    <tr><th> Type </th><td> float32 </td><td> numpy.ndarray </td></tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</td>\n",
       "<td>\n",
       "<svg width=\"170\" height=\"75\" style=\"stroke:rgb(0,0,0);stroke-width:1\" >\n",
       "\n",
       "  <!-- Horizontal lines -->\n",
       "  <line x1=\"0\" y1=\"0\" x2=\"120\" y2=\"0\" style=\"stroke-width:2\" />\n",
       "  <line x1=\"0\" y1=\"25\" x2=\"120\" y2=\"25\" style=\"stroke-width:2\" />\n",
       "\n",
       "  <!-- Vertical lines -->\n",
       "  <line x1=\"0\" y1=\"0\" x2=\"0\" y2=\"25\" style=\"stroke-width:2\" />\n",
       "  <line x1=\"6\" y1=\"0\" x2=\"6\" y2=\"25\" />\n",
       "  <line x1=\"12\" y1=\"0\" x2=\"12\" y2=\"25\" />\n",
       "  <line x1=\"18\" y1=\"0\" x2=\"18\" y2=\"25\" />\n",
       "  <line x1=\"25\" y1=\"0\" x2=\"25\" y2=\"25\" />\n",
       "  <line x1=\"31\" y1=\"0\" x2=\"31\" y2=\"25\" />\n",
       "  <line x1=\"37\" y1=\"0\" x2=\"37\" y2=\"25\" />\n",
       "  <line x1=\"44\" y1=\"0\" x2=\"44\" y2=\"25\" />\n",
       "  <line x1=\"50\" y1=\"0\" x2=\"50\" y2=\"25\" />\n",
       "  <line x1=\"56\" y1=\"0\" x2=\"56\" y2=\"25\" />\n",
       "  <line x1=\"63\" y1=\"0\" x2=\"63\" y2=\"25\" />\n",
       "  <line x1=\"69\" y1=\"0\" x2=\"69\" y2=\"25\" />\n",
       "  <line x1=\"75\" y1=\"0\" x2=\"75\" y2=\"25\" />\n",
       "  <line x1=\"82\" y1=\"0\" x2=\"82\" y2=\"25\" />\n",
       "  <line x1=\"88\" y1=\"0\" x2=\"88\" y2=\"25\" />\n",
       "  <line x1=\"94\" y1=\"0\" x2=\"94\" y2=\"25\" />\n",
       "  <line x1=\"101\" y1=\"0\" x2=\"101\" y2=\"25\" />\n",
       "  <line x1=\"107\" y1=\"0\" x2=\"107\" y2=\"25\" />\n",
       "  <line x1=\"113\" y1=\"0\" x2=\"113\" y2=\"25\" />\n",
       "  <line x1=\"120\" y1=\"0\" x2=\"120\" y2=\"25\" style=\"stroke-width:2\" />\n",
       "\n",
       "  <!-- Colored Rectangle -->\n",
       "  <polygon points=\"0.0,0.0 120.0,0.0 120.0,25.412616514582485 0.0,25.412616514582485\" style=\"fill:#8B4903A0;stroke-width:0\"/>\n",
       "\n",
       "  <!-- Text -->\n",
       "  <text x=\"60.000000\" y=\"45.412617\" font-size=\"1.0rem\" font-weight=\"100\" text-anchor=\"middle\" >1000000000</text>\n",
       "  <text x=\"140.000000\" y=\"12.706308\" font-size=\"1.0rem\" font-weight=\"100\" text-anchor=\"middle\" transform=\"rotate(0,140.000000,12.706308)\">1</text>\n",
       "</svg>\n",
       "</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "dask.array<array, shape=(1000000000,), dtype=float32, chunksize=(1000000,), chunktype=numpy.ndarray>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import dask.array as da\n",
    "x = da.from_array(dset, chunks=(1_000_000,))\n",
    "x"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37e12bf0-7b72-467b-b7bf-9dd9e3ad2620",
   "metadata": {},
   "source": [
    "Now that we have an `Array` we perform standard numpy-style computations like arithmetic, mathematics, slicing, reductions, etc..\n",
    "\n",
    "The interface is familiar, but the actual work is different. `dask_array.sum()` does not do the same thing as `numpy_array.sum()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "512ddf0f-9ae0-40a6-b0db-6cf7842ff507",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2.96 s, sys: 113 ms, total: 3.07 s\n",
      "Wall time: 5.46 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1000010200.0"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "result = x.sum()\n",
    "result.compute()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "1b343430-a6ab-41bc-80f5-ba1923de17cf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.84 s, sys: 61.5 ms, total: 1.9 s\n",
      "Wall time: 3.03 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1.0000101"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "result = x.mean()\n",
    "result.compute()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c23e531-8135-4207-b7cd-2f5c53e22ba6",
   "metadata": {},
   "source": [
    "### Performance and Parallelism\n",
    "\n",
    "<img src=\"assets/fail-case.gif\" width=\"40%\" align=\"right\">\n",
    "\n",
    "In our first examples we used `for` loops to walk through the array one block at a time.  For simple operations like `sum` this is optimal.  However for complex operations we may want to traverse through the array differently.  In particular we may want the following:\n",
    "\n",
    "1.  Use multiple cores in parallel\n",
    "2.  Chain operations on a single blocks before moving on to the next one\n",
    "\n",
    "`Dask.array` translates your array operations into a graph of inter-related tasks with data dependencies between them. Dask then executes this graph in parallel with multiple threads.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74cf728f-ff75-4fca-8c3f-417313db22a5",
   "metadata": {},
   "source": [
    "#### Example\n",
    "\n",
    "1.  Construct a 20000x20000 array of normally distributed random values broken up into 1000x1000 sized chunks\n",
    "2.  Take the mean along one axis\n",
    "3.  Take every 100th element"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ddce129d-0a5d-44b7-bad6-238cd3f8ec24",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import dask.array as da\n",
    "\n",
    "x = da.random.normal(10, 0.1, size=(20000, 20000),   # 400 million element array \n",
    "                              chunks=(1000, 1000))   # Cut into 1000x1000 sized chunks\n",
    "y = x.mean(axis=0)[::100]                            # Perform NumPy-style operations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "dabd7be0-743b-4b21-8f6e-16ade8e77cc8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.2"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.nbytes / 1e9  # Gigabytes of the input processed lazily"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "131974cc-1035-4e0a-91f6-ca4aa7980e66",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.13 s, sys: 38.2 ms, total: 1.17 s\n",
      "Wall time: 1.34 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 9.9995448 ,  9.99996055,  9.99898756,  9.99894536,  9.99971163,\n",
       "       10.00014634, 10.00023615,  9.99867533, 10.00057959, 10.00013251,\n",
       "        9.99877735,  9.99975823, 10.00035141, 10.00111201, 10.00095824,\n",
       "       10.0001436 ,  9.9992607 , 10.00019998,  9.99918585, 10.00027708,\n",
       "       10.0014034 ,  9.99993418, 10.00058631, 10.00008997,  9.999656  ,\n",
       "        9.99987386, 10.00010253,  9.99960963,  9.99961663,  9.99964701,\n",
       "       10.00105797,  9.9991494 ,  9.99966805, 10.00022704,  9.99976607,\n",
       "        9.99990729,  9.99876798,  9.99987097, 10.00015794, 10.00062863,\n",
       "       10.00023233, 10.00006062,  9.99959999,  9.99951237, 10.00006846,\n",
       "       10.00013044,  9.99994292, 10.00057899,  9.99995689, 10.00047121,\n",
       "       10.00000773,  9.99892913, 10.00047937,  9.99994692, 10.00043925,\n",
       "       10.00017603,  9.99968372, 10.00024172,  9.9992128 , 10.00068941,\n",
       "        9.99938285,  9.99959026, 10.00016617, 10.00125664,  9.99942822,\n",
       "        9.99973289, 10.00048579,  9.99980449, 10.00017378, 10.00018803,\n",
       "        9.99982612, 10.00011156, 10.0003825 ,  9.99974018, 10.00042776,\n",
       "        9.99923848, 10.00078315, 10.00021434,  9.99912496,  9.99980412,\n",
       "       10.00145747,  9.99945282, 10.00111557,  9.99965383,  9.9994168 ,\n",
       "        9.99962983,  9.99829098,  9.9991091 , 10.00023444,  9.99961168,\n",
       "       10.00111302,  9.99984671,  9.99997567,  9.99974406,  9.99850961,\n",
       "        9.99993162, 10.00142668, 10.00049849,  9.99940278, 10.000666  ,\n",
       "       10.00004627,  9.99998838, 10.00110967, 10.00041693, 10.0009037 ,\n",
       "        9.9994393 , 10.00005653,  9.99885879,  9.99958346,  9.99981265,\n",
       "       10.00032571,  9.99956745,  9.99957837,  9.99983736, 10.00027929,\n",
       "       10.00011384, 10.00056338, 10.0003938 , 10.00112545,  9.99977495,\n",
       "       10.00010663, 10.00062939, 10.00093876,  9.99980097,  9.99941842,\n",
       "        9.99942856, 10.00081627,  9.99981516, 10.00092848, 10.00085166,\n",
       "       10.0002889 , 10.00017003,  9.99916245, 10.00099707, 10.00009261,\n",
       "       10.00016014,  9.99975552, 10.00007985, 10.00012469, 10.00122204,\n",
       "        9.99976303, 10.00103567, 10.00024406, 10.00026369,  9.99888962,\n",
       "       10.00001441, 10.00034358, 10.00015895, 10.00146557,  9.99947387,\n",
       "        9.99906356,  9.99960949, 10.00054188,  9.99859927,  9.99903953,\n",
       "       10.00031349, 10.00086996, 10.0003289 , 10.00020636,  9.99894613,\n",
       "        9.99921386, 10.00151932,  9.99979472,  9.9991016 , 10.00062258,\n",
       "       10.00046631,  9.99986447,  9.9994954 ,  9.99961436, 10.00064818,\n",
       "        9.99892503,  9.99964759, 10.00027975, 10.00129814, 10.00018721,\n",
       "        9.99878972, 10.00014764,  9.9994935 ,  9.99994014, 10.00103056,\n",
       "       10.00024078, 10.00143084, 10.00029025,  9.99954135, 10.00027938,\n",
       "       10.00113807, 10.00055951, 10.00069637, 10.00136242,  9.99971921,\n",
       "       10.00014468,  9.99945525,  9.99994246, 10.00071985, 10.00088819,\n",
       "        9.99962296, 10.00131722,  9.99979625,  9.99996547, 10.00024176])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "y.compute()     # Time to compute the result"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79b59faa-7362-4e99-a3ad-d1325233f49b",
   "metadata": {},
   "source": [
    "#### NumPy: 19s, Needs gigabytes of memory\n",
    "\n",
    "```python\n",
    "import numpy as np\n",
    "\n",
    "%%time \n",
    "x = np.random.normal(10, 0.1, size=(20000, 20000)) \n",
    "y = x.mean(axis=0)[::100] \n",
    "y\n",
    "\n",
    "CPU times: user 19.6 s, sys: 160 ms, total: 19.8 s\n",
    "Wall time: 19.7 s\n",
    "```\n",
    "\n",
    "#### Dask Array: 4s, Needs megabytes of memory\n",
    "\n",
    "```python\n",
    "import dask.array as da\n",
    "\n",
    "%%time\n",
    "x = da.random.normal(10, 0.1, size=(20000, 20000), chunks=(1000, 1000))\n",
    "y = x.mean(axis=0)[::100] \n",
    "y.compute() \n",
    "\n",
    "CPU times: user 29.4 s, sys: 1.07 s, total: 30.5 s\n",
    "Wall time: 4.01 s\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "e3e27094-881f-49eb-be51-9cf8e93d7f5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "client.shutdown()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3de2c22c-af15-4334-91ef-d094507ac256",
   "metadata": {},
   "source": [
    "# Dataframes: analagous to Pandas DataFrames\n",
    "\n",
    "<img src=\"assets/dask-dataframe.png\" width=\"50%\">\n",
    "\n",
    "In this section we use `dask.dataframe` to automatically build parallel dataframe computations over a directory of CSV files, for the common case of tabular computations.  Dask dataframes look and feel like Pandas dataframes but they run on the same infrastructure that powers `dask.delayed`.\n",
    "\n",
    "Here, we'll create airline data, and we'll let `dask.dataframe` construct our computations for us.  The `dask.dataframe.read_csv` function can take a globstring like `\"data/nycflights/*.csv\"` and build parallel computations on all of our data at once.\n",
    "\n",
    "### When to use `dask.dataframe`\n",
    "\n",
    "Pandas is great for tabular datasets that fit in memory. Dask becomes useful when the dataset you want to analyze is larger than your machine's RAM. The demo dataset we're working with is only about 200MB, so that you can download it in a reasonable time, but `dask.dataframe` will scale to  datasets much larger than memory."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6aa02279-2735-41ed-863b-51d3b99dc2c4",
   "metadata": {},
   "source": [
    "<img src=\"assets/pandas_logo.png\" align=\"right\" width=\"28%\">\n",
    "\n",
    "The `dask.dataframe` module implements a blocked parallel `DataFrame` object that mimics a large subset of the Pandas `DataFrame` API. One Dask `DataFrame` is comprised of many in-memory pandas `DataFrames` separated along the index. One operation on a Dask `DataFrame` triggers many pandas operations on the constituent pandas `DataFrame`s in a way that is mindful of potential parallelism and memory constraints.\n",
    "\n",
    "**Related Documentation**\n",
    "\n",
    "* [DataFrame documentation](https://docs.dask.org/en/latest/dataframe.html)\n",
    "* [DataFrame screencast](https://youtu.be/AT2XtFehFSQ)\n",
    "* [DataFrame API](https://docs.dask.org/en/latest/dataframe-api.html)\n",
    "* [DataFrame examples](https://examples.dask.org/dataframe.html)\n",
    "* [Pandas documentation](https://pandas.pydata.org/pandas-docs/stable/)\n",
    "\n",
    "**Main Take-aways**\n",
    "\n",
    "1.  Dask DataFrame should be familiar to Pandas users\n",
    "2.  The partitioning of dataframes is important for efficient execution"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e6b9b16-e764-45c9-b47c-a67ac4cddf46",
   "metadata": {},
   "source": [
    "### Create Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "44fab43f-bfd9-4fa9-942c-2b225f3c2552",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "- Downloading NYC Flights dataset... "
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "distributed.client - ERROR - Failed to reconnect to scheduler after 10.00 seconds, closing client\n",
      "_GatheringFuture exception was never retrieved\n",
      "future: <_GatheringFuture finished exception=CancelledError()>\n",
      "asyncio.exceptions.CancelledError\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done\n",
      "- Extracting flight data... done\n",
      "- Creating json data... done\n",
      "** Created flights dataset! in 18.56s**\n"
     ]
    }
   ],
   "source": [
    "%run prep.py -d flights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "d4b9b0e4-6ffc-431a-ab67-7ab9eaa8edb9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dask.distributed import Client\n",
    "\n",
    "client = Client(n_workers=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f6a2b75-350f-4078-af5d-b21f91690eb1",
   "metadata": {},
   "source": [
    "### Reading and Viewing Dask DataFrames\n",
    "\n",
    "Let's try to analyze an extract of flights in the USA across several years. This data is specific to flights out of the three airports in the New York City area."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "7f5b5994-82ed-467a-9413-97b7499a4d79",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><strong>Dask DataFrame Structure:</strong></div>\n",
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>DepTime</th>\n",
       "      <th>CRSDepTime</th>\n",
       "      <th>ArrTime</th>\n",
       "      <th>CRSArrTime</th>\n",
       "      <th>UniqueCarrier</th>\n",
       "      <th>FlightNum</th>\n",
       "      <th>TailNum</th>\n",
       "      <th>ActualElapsedTime</th>\n",
       "      <th>CRSElapsedTime</th>\n",
       "      <th>AirTime</th>\n",
       "      <th>ArrDelay</th>\n",
       "      <th>DepDelay</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Dest</th>\n",
       "      <th>Distance</th>\n",
       "      <th>TaxiIn</th>\n",
       "      <th>TaxiOut</th>\n",
       "      <th>Cancelled</th>\n",
       "      <th>Diverted</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>npartitions=10</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <td>datetime64[ns]</td>\n",
       "      <td>int64</td>\n",
       "      <td>float64</td>\n",
       "      <td>int64</td>\n",
       "      <td>float64</td>\n",
       "      <td>int64</td>\n",
       "      <td>object</td>\n",
       "      <td>int64</td>\n",
       "      <td>object</td>\n",
       "      <td>float64</td>\n",
       "      <td>float64</td>\n",
       "      <td>float64</td>\n",
       "      <td>float64</td>\n",
       "      <td>float64</td>\n",
       "      <td>object</td>\n",
       "      <td>object</td>\n",
       "      <td>float64</td>\n",
       "      <td>float64</td>\n",
       "      <td>float64</td>\n",
       "      <td>bool</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "<div>Dask Name: read-csv, 10 tasks</div>"
      ],
      "text/plain": [
       "Dask DataFrame Structure:\n",
       "                          Date DayOfWeek  DepTime CRSDepTime  ArrTime CRSArrTime UniqueCarrier FlightNum TailNum ActualElapsedTime CRSElapsedTime  AirTime ArrDelay DepDelay  Origin    Dest Distance   TaxiIn  TaxiOut Cancelled Diverted\n",
       "npartitions=10                                                                                                                                                                                                                            \n",
       "                datetime64[ns]     int64  float64      int64  float64      int64        object     int64  object           float64        float64  float64  float64  float64  object  object  float64  float64  float64      bool    int64\n",
       "                           ...       ...      ...        ...      ...        ...           ...       ...     ...               ...            ...      ...      ...      ...     ...     ...      ...      ...      ...       ...      ...\n",
       "...                        ...       ...      ...        ...      ...        ...           ...       ...     ...               ...            ...      ...      ...      ...     ...     ...      ...      ...      ...       ...      ...\n",
       "                           ...       ...      ...        ...      ...        ...           ...       ...     ...               ...            ...      ...      ...      ...     ...     ...      ...      ...      ...       ...      ...\n",
       "                           ...       ...      ...        ...      ...        ...           ...       ...     ...               ...            ...      ...      ...      ...     ...     ...      ...      ...      ...       ...      ...\n",
       "Dask Name: read-csv, 10 tasks"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import dask.dataframe as dd\n",
    "\n",
    "df = dd.read_csv(os.path.join('data', 'nycflights', '*.csv'),\n",
    "                 parse_dates={'Date': [0, 1, 2]},\n",
    "                 dtype={'TailNum': str,\n",
    "                        'CRSElapsedTime': float,\n",
    "                        'Cancelled': bool})\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "be742463-3065-4da7-aafa-0ea884682f2a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>DepTime</th>\n",
       "      <th>CRSDepTime</th>\n",
       "      <th>ArrTime</th>\n",
       "      <th>CRSArrTime</th>\n",
       "      <th>UniqueCarrier</th>\n",
       "      <th>FlightNum</th>\n",
       "      <th>TailNum</th>\n",
       "      <th>ActualElapsedTime</th>\n",
       "      <th>...</th>\n",
       "      <th>AirTime</th>\n",
       "      <th>ArrDelay</th>\n",
       "      <th>DepDelay</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Dest</th>\n",
       "      <th>Distance</th>\n",
       "      <th>TaxiIn</th>\n",
       "      <th>TaxiOut</th>\n",
       "      <th>Cancelled</th>\n",
       "      <th>Diverted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1990-01-01</td>\n",
       "      <td>1</td>\n",
       "      <td>1621.0</td>\n",
       "      <td>1540</td>\n",
       "      <td>1747.0</td>\n",
       "      <td>1701</td>\n",
       "      <td>US</td>\n",
       "      <td>33</td>\n",
       "      <td>NaN</td>\n",
       "      <td>86.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>46.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>EWR</td>\n",
       "      <td>PIT</td>\n",
       "      <td>319.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1990-01-02</td>\n",
       "      <td>2</td>\n",
       "      <td>1547.0</td>\n",
       "      <td>1540</td>\n",
       "      <td>1700.0</td>\n",
       "      <td>1701</td>\n",
       "      <td>US</td>\n",
       "      <td>33</td>\n",
       "      <td>NaN</td>\n",
       "      <td>73.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>EWR</td>\n",
       "      <td>PIT</td>\n",
       "      <td>319.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1990-01-03</td>\n",
       "      <td>3</td>\n",
       "      <td>1546.0</td>\n",
       "      <td>1540</td>\n",
       "      <td>1710.0</td>\n",
       "      <td>1701</td>\n",
       "      <td>US</td>\n",
       "      <td>33</td>\n",
       "      <td>NaN</td>\n",
       "      <td>84.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>EWR</td>\n",
       "      <td>PIT</td>\n",
       "      <td>319.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1990-01-04</td>\n",
       "      <td>4</td>\n",
       "      <td>1542.0</td>\n",
       "      <td>1540</td>\n",
       "      <td>1710.0</td>\n",
       "      <td>1701</td>\n",
       "      <td>US</td>\n",
       "      <td>33</td>\n",
       "      <td>NaN</td>\n",
       "      <td>88.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>EWR</td>\n",
       "      <td>PIT</td>\n",
       "      <td>319.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1990-01-05</td>\n",
       "      <td>5</td>\n",
       "      <td>1549.0</td>\n",
       "      <td>1540</td>\n",
       "      <td>1706.0</td>\n",
       "      <td>1701</td>\n",
       "      <td>US</td>\n",
       "      <td>33</td>\n",
       "      <td>NaN</td>\n",
       "      <td>77.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>EWR</td>\n",
       "      <td>PIT</td>\n",
       "      <td>319.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date  DayOfWeek  DepTime  CRSDepTime  ArrTime  CRSArrTime  \\\n",
       "0 1990-01-01          1   1621.0        1540   1747.0        1701   \n",
       "1 1990-01-02          2   1547.0        1540   1700.0        1701   \n",
       "2 1990-01-03          3   1546.0        1540   1710.0        1701   \n",
       "3 1990-01-04          4   1542.0        1540   1710.0        1701   \n",
       "4 1990-01-05          5   1549.0        1540   1706.0        1701   \n",
       "\n",
       "  UniqueCarrier  FlightNum TailNum  ActualElapsedTime  ...  AirTime  ArrDelay  \\\n",
       "0            US         33     NaN               86.0  ...      NaN      46.0   \n",
       "1            US         33     NaN               73.0  ...      NaN      -1.0   \n",
       "2            US         33     NaN               84.0  ...      NaN       9.0   \n",
       "3            US         33     NaN               88.0  ...      NaN       9.0   \n",
       "4            US         33     NaN               77.0  ...      NaN       5.0   \n",
       "\n",
       "   DepDelay  Origin Dest Distance  TaxiIn  TaxiOut  Cancelled  Diverted  \n",
       "0      41.0     EWR  PIT    319.0     NaN      NaN      False         0  \n",
       "1       7.0     EWR  PIT    319.0     NaN      NaN      False         0  \n",
       "2       6.0     EWR  PIT    319.0     NaN      NaN      False         0  \n",
       "3       2.0     EWR  PIT    319.0     NaN      NaN      False         0  \n",
       "4       9.0     EWR  PIT    319.0     NaN      NaN      False         0  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88453d9f-f358-4a4c-9e86-09aab6a66acb",
   "metadata": {},
   "source": [
    "### Computations with `dask.dataframe`\n",
    "\n",
    "We compute the maximum of the `DepDelay` column. With just pandas, we would loop over each file to find the individual maximums, then find the final maximum over all the individual maximums\n",
    "\n",
    "```python\n",
    "maxes = []\n",
    "for fn in filenames:\n",
    "    df = pd.read_csv(fn)\n",
    "    maxes.append(df.DepDelay.max())\n",
    "    \n",
    "final_max = max(maxes)\n",
    "```\n",
    "\n",
    "We could wrap that `pd.read_csv` with `dask.delayed` so that it runs in parallel. Regardless, we're still having to think about loops, intermediate results (one per file) and the final reduction (`max` of the intermediate maxes). This is just noise around the real task, which pandas solves with\n",
    "\n",
    "```python\n",
    "df = pd.read_csv(filename, dtype=dtype)\n",
    "df.DepDelay.max()\n",
    "```\n",
    "\n",
    "`dask.dataframe` lets us write pandas-like code, that operates on larger than memory datasets in parallel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "b1435054-91d9-4b67-a470-4c392f409101",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.76 s, sys: 70.8 ms, total: 1.83 s\n",
      "Wall time: 4.4 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1435.0"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "df['DepDelay'].max().compute()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "ad454526-4662-4ca0-a5c2-b3bb9e83e1a0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 767 ms, sys: 32.4 ms, total: 800 ms\n",
      "Wall time: 1.64 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Origin\n",
       "EWR    10.295469\n",
       "JFK    10.351299\n",
       "LGA     7.431142\n",
       "Name: DepDelay, dtype: float64"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "df.groupby('Origin')['DepDelay'].mean().compute()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "94610249-ec2d-47a3-b695-28d3e2770875",
   "metadata": {},
   "outputs": [],
   "source": [
    "client.shutdown()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b478db19-9df5-4588-9470-d26600b7ce3f",
   "metadata": {},
   "source": [
    "# Futures interface\n",
    "\n",
    "The Dask distributed scheduler implements a superset of Python's [`concurrent.futures`](https://docs.python.org/3/library/concurrent.futures.html) interface that allows for finer control and asynchronous computation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "46cb5d57-a0ba-4c84-b044-325dae9d49b8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table style=\"border: 2px solid white;\">\n",
       "<tr>\n",
       "<td style=\"vertical-align: top; border: 0px solid white\">\n",
       "<h3 style=\"text-align: left;\">Client</h3>\n",
       "<ul style=\"text-align: left; list-style: none; margin: 0; padding: 0;\">\n",
       "  <li><b>Scheduler: </b>tcp://127.0.0.1:62595</li>\n",
       "  <li><b>Dashboard: </b><a href='http://127.0.0.1:8787/status' target='_blank'>http://127.0.0.1:8787/status</a></li>\n",
       "</ul>\n",
       "</td>\n",
       "<td style=\"vertical-align: top; border: 0px solid white\">\n",
       "<h3 style=\"text-align: left;\">Cluster</h3>\n",
       "<ul style=\"text-align: left; list-style:none; margin: 0; padding: 0;\">\n",
       "  <li><b>Workers: </b>4</li>\n",
       "  <li><b>Cores: </b>16</li>\n",
       "  <li><b>Memory: </b>16.00 GiB</li>\n",
       "</ul>\n",
       "</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<Client: 'tcp://127.0.0.1:62595' processes=4 threads=16, memory=16.00 GiB>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "distributed.client - ERROR - Failed to reconnect to scheduler after 10.00 seconds, closing client\n",
      "_GatheringFuture exception was never retrieved\n",
      "future: <_GatheringFuture finished exception=CancelledError()>\n",
      "asyncio.exceptions.CancelledError\n"
     ]
    }
   ],
   "source": [
    "from dask.distributed import Client\n",
    "\n",
    "client = Client(n_workers=4)\n",
    "client"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "8bc1791e-43d8-48e3-8717-f168cc9944ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "import random\n",
    "\n",
    "def inc(x):\n",
    "    time.sleep(random.random())\n",
    "    return x + 1\n",
    "\n",
    "def double(x):\n",
    "    time.sleep(random.random())\n",
    "    return 2 * x\n",
    "    \n",
    "def add(x, y):\n",
    "    time.sleep(random.random())\n",
    "    return x + y "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec3da8f0-70f8-4339-a1ee-cabfb1ad0dd4",
   "metadata": {},
   "source": [
    "We can run these functions locally"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "581aab70-c0ad-406b-a656-cf903b35ac4f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inc(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9fd7af10-a0da-4713-806f-147dace34f8a",
   "metadata": {},
   "source": [
    "Or we can submit them to run remotely on a Dask worker node"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "60a9be06-537c-462d-85f4-1e3847a53cff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<b>Future: inc</b> <font color=\"gray\">status: </font><font color=\"black\">pending</font>, <font color=\"gray\">key: </font>inc-8e0a31564998e7d9ef3ab85b18ffa1c3"
      ],
      "text/plain": [
       "<Future: pending, key: inc-8e0a31564998e7d9ef3ab85b18ffa1c3>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future = client.submit(inc, 1)\n",
    "future"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c5b3b60a-3bbf-4f83-a5ee-6efa2fcf0273",
   "metadata": {},
   "source": [
    "The `submit` function sends a function and arguments to the distributed scheduler for processing. It returns a `Future` object that refer to remote data on the cluster. The `Future` returns immediately while the computations run remotely in the background. There is no blocking of the local Python session.\n",
    "\n",
    "If you wait a moment, and then check on the future again, you'll see that it has finished."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "ca5816bc-03ed-4625-8576-db805bbdd824",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<b>Future: inc</b> <font color=\"gray\">status: </font><font color=\"black\">finished</font>, <font color=\"gray\">type: </font>builtins.int, <font color=\"gray\">key: </font>inc-8e0a31564998e7d9ef3ab85b18ffa1c3"
      ],
      "text/plain": [
       "<Future: finished, type: builtins.int, key: inc-8e0a31564998e7d9ef3ab85b18ffa1c3>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0135a363-df2a-490b-baf0-4f30e0baeab7",
   "metadata": {},
   "source": [
    "Once the computation for a `Future` is complete, you can retrieve the result using the `.result()` method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "450b2f6f-ec57-490b-9ea2-bab9fe1c1115",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future.result()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd4b5045-fa34-47cd-b47f-be05e22406bf",
   "metadata": {},
   "source": [
    "### Specifying task dependencies\n",
    "\n",
    "Much like the `delayed` interface, we can submit tasks based on other futures. This will create a dependency between the inputs and outputs. Dask will track the execution of all tasks and ensure that downstream tasks are run at the proper time and place and with the proper data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "2c3b63f9-9062-4fbb-b827-10c57c0d2c84",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<b>Future: add</b> <font color=\"gray\">status: </font><font color=\"black\">pending</font>, <font color=\"gray\">key: </font>add-f90c64be930e9620aeacd6fca2b9a5d6"
      ],
      "text/plain": [
       "<Future: pending, key: add-f90c64be930e9620aeacd6fca2b9a5d6>"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = client.submit(inc, 1)\n",
    "y = client.submit(double, 2)\n",
    "z = client.submit(add, x, y)\n",
    "z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "6b79e47f-5e1b-4ff0-aa1a-47bc084c992c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "z.result()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee28e621-d6bb-45eb-986d-0881a2e44822",
   "metadata": {},
   "source": [
    "As an example, we can submit many tasks that depend on each other in a `for`-loop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "ce9ee845-3fd4-41f5-be6e-fa11250a8318",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 97.4 ms, sys: 7.25 ms, total: 105 ms\n",
      "Wall time: 100 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "zs = []\n",
    "\n",
    "for i in range(64):\n",
    "    x = client.submit(inc, i)     # x = inc(i)\n",
    "    y = client.submit(double, x)  # y = inc(x)\n",
    "    z = client.submit(add, x, y)  # z = inc(y)\n",
    "    zs.append(z)\n",
    "\n",
    "total = client.submit(sum, zs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "b583c90d-d0ce-4e79-9132-8a322322494d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<b>Future: sum</b> <font color=\"gray\">status: </font><font color=\"black\">finished</font>, <font color=\"gray\">type: </font>builtins.int, <font color=\"gray\">key: </font>sum-9963cdfa333d82f2e4a90f4fac4df0d1"
      ],
      "text/plain": [
       "<Future: finished, type: builtins.int, key: sum-9963cdfa333d82f2e4a90f4fac4df0d1>"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "dbc4a11e-a90f-4272-ae8e-a5b6343ab5c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6240"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total.result()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9649cb92-ad88-41cf-822d-f0207a99ecae",
   "metadata": {},
   "outputs": [],
   "source": [
    "client.shutdown()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}