Visualizations.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "import collections\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.ticker\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import glob\n",
    "import re\n",
    "import json\n",
    "from scipy import stats\n",
    "import os\n",
    "import itertools\n",
    "import bisect\n",
    "\n",
    "chunk_sizes_ = [63171, 65535, 4427, 65535, 4702, 65535, 5149, 65535, 4768, 65535, 4351, 65535, 4875, 65535, 5292, 65535, 2276, 65535, 2602, 65535, 2747, 65535, 2600, 65535, 2566, 65535, 2200, 65535, 3376, 65535, 2779, 65535, 2176, 65535, 3521, 65535, 2083, 65535, 2528, 65535, 2168, 65535, 2658, 65535, 3041, 65535, 3120, 65535, 2558, 65535, 2717, 65535, 2582, 65535, 2134, 65535, 2388, 65535, 2755, 65535, 2688, 59138, 63841, 65535, 4586, 65535, 4685, 65535, 4736, 65535, 5099, 65535, 4831, 65535, 4942, 65535, 5521, 65535, 2339, 65535, 2495, 65535, 2596, 65535, 2571, 65535, 2735, 65535, 2455, 65535, 2836, 65535, 3141, 65535, 3066, 65535, 3156, 65535, 2495, 65535, 1844, 65535, 2313, 65535, 2612, 65535, 3017, 65535, 2911, 65535, 2377, 65535, 2841, 65535, 2903, 65535, 2016, 65535, 2299, 65535, 2496, 65535, 2343, 58954, 47296, 52954, 52703, 52389, 53230, 53110, 53276, 52886, 50913, 50895, 51180, 50548, 50525, 50911, 51196, 51408, 51118, 50985, 50788, 50915, 51129, 51033, 51321, 50903, 51183, 50824, 51240, 51194, 51245, 51042, 50915, 44672,]\n",
    "chunk_sizes_ = [63171, 63841, 47296, 65535, 4427, 65535, 4586, 52954, 65535, 4702, 65535, 4685, 52703, 65535, 5149, 65535, 4736, 52389, 65535, 4768, 65535, 5099, 53230, 65535, 4351, 65535, 4831, 53110, 65535, 4875, 65535, 4942, 53276, 65535, 5292, 65535, 5521, 52886, 65535, 65535, 50913, 65535, 65535, 50895, 65535, 65535, 51180, 65535, 65535, 50548, 65535, 65535, 50525, 65535, 65535, 50911, 65535, 65535, 51196, 65535, 65535, 51408, 65535, 65535, 51118, 65535, 65535, 50985, 65535, 65535, 50788, 65535, 65535, 50915, 65535, 65535, 51129, 65535, 65535, 51033, 65535, 65535, 51321, 65535, 65535, 50903, 65535, 65535, 51183, 65535, 65535, 50824, 65535, 65535, 51240, 65535, 65535, 51194, 65535, 65535, 51245, 65535, 65535, 51042, 65535, 65535, 50915, 59138, 58954, 44672, 65535, 54585, ]\n",
    "chunk_sizes = [63171, 63841, 47296, 65535, 65535, 52954, 65535, 65535, 52703, 65535, 65535, 52389, 65535, 65535, 53230, 65535, 65535, 53110, 65535, 65535, 53276, 65535, 65535, 52886, 65535, 65535, 50913, 65535, 65535, 50895, 65535, 65535, 51180, 65535, 65535, 50548, 65535, 65535, 50525, 65535, 65535, 50911, 65535, 65535, 51196, 65535, 65535, 51408, 65535, 65535, 51118, 65535, 65535, 50985, 65535, 65535, 50788, 65535, 65535, 50915, 65535, 65535, 51129, 65535, 65535, 51033, 65535, 65535, 51321, 65535, 65535, 50903, 65535, 65535, 51183, 65535, 65535, 50824, 65535, 65535, 51240, 65535, 65535, 51194, 65535, 65535, 51245, 65535, 65535, 51042, 65535, 65535, 50915, 59138, 58954, 44672, 65535, 65535, 57014, ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0, 0, 0, 0, 0, 0, 0, 0, 2, 30, 5, 62]\n"
     ]
    }
   ],
   "source": [
    "bins = [0, 100, 500, 1000, 2000, 4000, 8000, 16000, 32000, 48000, 55000, 65535]\n",
    "\n",
    "def chunks_with_x_elements(low, high, upper_exclusive=True):\n",
    "    if upper_exclusive:        \n",
    "        filtered = filter(lambda x: low <= x and x < high, chunk_sizes)\n",
    "    else:\n",
    "        filtered = filter(lambda x: low <= x and x <= high, chunk_sizes)\n",
    "            \n",
    "    return len(list(filtered))\n",
    "\n",
    "chunks_per_bin = [chunks_with_x_elements(low, high ) for low, high in zip(bins, bins[1:])]\n",
    "chunks_per_bin.append(chunks_with_x_elements(65535, 65535, False))\n",
    "\n",
    "\n",
    "labels_per_bin = [str(low) + \"-\" + str(high-1) for low, high in zip(bins, bins[1:])]\n",
    "labels_per_bin.append(str(65535))\n",
    "\n",
    "print(chunks_per_bin)\n",
    "\n",
    "assert sum(chunks_per_bin) == len(chunk_sizes), \"Bug\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_bins = np.arange(len(bins))\n",
    "\n",
    "plt.bar(num_bins, chunks_per_bin)\n",
    "plt.title(\"Number of chunks with x rows\")\n",
    "plt.xlabel(\"Rows\")\n",
    "plt.ylabel(\"Chunks\")\n",
    "plt.xticks(num_bins, labels_per_bin, rotation='vertical')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "distinct_values = [65535/4*1, 65535/4*2, 65535/4*3, 65535/4*4]\n",
    "yerr = np.arange(1, 5) * 2\n",
    "\n",
    "runtimes = [113, 128, 134, 175]\n",
    "plt.errorbar(distinct_values, runtimes, yerr=yerr, label='Unsorted chunks')\n",
    "\n",
    "runtimes_sorted = [72, 80, 90, 105]\n",
    "plt.errorbar(distinct_values, runtimes_sorted, yerr=yerr, label='Sorted Chunks')\n",
    "\n",
    "runtimes_sorted_influence = np.array(runtimes) / np.array(runtimes_sorted) * 100\n",
    "plt.errorbar(distinct_values, runtimes_sorted_influence, yerr=yerr, label='Unsorted / Sorted * 100')\n",
    "\n",
    "#runtimes_mixed = [94, 113, 132, 151, 170]\n",
    "#plt.errorbar([0]+distinct_values, runtimes_mixed, yerr=[1,2,3,4,5])\n",
    "\n",
    "#speedups = np.array([175/113, 175/128, 175/134, 175/175]) * 100\n",
    "#plt.errorbar(distinct_values, speedups, yerr=yerr)\n",
    "\n",
    "#speedups_guess = np.array([175/95, 175/175]) * 100\n",
    "#plt.errorbar([0, 65535], speedups_guess)\n",
    "\n",
    "plt.xlim(xmin=0)\n",
    "plt.xlabel(\"Disjoint values per chunk\")\n",
    "plt.ylim(ymin=0)\n",
    "plt.ylabel(\"Runtime Materialize-Step (JoinHash) in ms\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def align_yaxis(ax1, v1, ax2, v2):\n",
    "    \"\"\"adjust ax2 ylimit so that v2 in ax2 is aligned to v1 in ax1\"\"\"\n",
    "    _, y1 = ax1.transData.transform((0, v1))\n",
    "    _, y2 = ax2.transData.transform((0, v2))\n",
    "    adjust_yaxis(ax2,(y1-y2)/2,v2)\n",
    "    adjust_yaxis(ax1,(y2-y1)/2,v1)\n",
    "\n",
    "def adjust_yaxis(ax,ydif,v):\n",
    "    \"\"\"shift axis ax by ydiff, maintaining point v at the same location\"\"\"\n",
    "    inv = ax.transData.inverted()\n",
    "    _, dy = inv.transform((0, 0)) - inv.transform((0, ydif))\n",
    "    miny, maxy = ax.get_ylim()\n",
    "    miny, maxy = miny - v, maxy - v\n",
    "    if -miny>maxy or (-miny==maxy and dy > 0):\n",
    "        nminy = miny\n",
    "        nmaxy = miny*(maxy+dy)/(miny+dy)\n",
    "    else:\n",
    "        nmaxy = maxy\n",
    "        nminy = maxy*(miny+dy)/(maxy+dy)\n",
    "    ax.set_ylim(nminy+v, nmaxy+v)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "queries = [3, 4, 5, 6, 7, 9, 12, 14, 15, 18, 20, 21]\n",
    "queries = [str(x) for x in queries]\n",
    "throughput = np.array([25, -21, -14, 318, 38, -16, 68, 88, 146, -24, 166, -15])\n",
    "latency = [251, 132, 435, 35, 177, 948, 146, 80, 45, 998, 90, 725]\n",
    "\n",
    "x = np.arange(len(queries))  # the label locations\n",
    "width = 0.35  # the width of the bars\n",
    "\n",
    "assert len(queries) == len(throughput), \"Mismatch 1\"\n",
    "assert len(queries) == len(latency), \"Mismatch 2\"\n",
    "print(f\"Considering {len(queries)} queries\")\n",
    "\n",
    "fig, ax1 = plt.subplots()\n",
    "\n",
    "color = 'tab:red'\n",
    "ax1.set_xlabel('TPC-H Query')\n",
    "ax1.set_xticks(x)\n",
    "ax1.set_xticklabels(queries)\n",
    "ax1.set_ylabel('Change in %', color='black')\n",
    "throughput_bar = ax1.bar(x - width/2, throughput, width, label='Throughput', color=color)\n",
    "ax1.tick_params(axis='y', labelcolor='black')\n",
    "#ax1.legend()\n",
    "\n",
    "ax2 = ax1.twinx()  # instantiate a second axes that shares the same x-axis\n",
    "\n",
    "color = 'tab:blue'\n",
    "ax2.set_ylabel('Milliseconds', color='black')  # we already handled the x-label with ax1\n",
    "latency_bar = ax2.bar(x + width/2, latency, width, label='Latency', color=color)\n",
    "ax2.tick_params(axis='y', labelcolor='black')\n",
    "#ax2.legend()\n",
    "\n",
    "align_yaxis(ax1, 0, ax2, 0)\n",
    "\n",
    "plt.legend([throughput_bar, latency_bar], ['Throughput', 'Latency'])\n",
    "fig.tight_layout()  # otherwise the right y-label is slightly clipped\n",
    "plt.savefig('/home/aloeser/Downloads/throughput_latency.pdf')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32945.669379\n",
      "33336.943587\n",
      "347\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Graph for the total runtime per hash join step\n",
    "BENCHMARK = \"TPC-DS\"\n",
    "if BENCHMARK == \"TPC-H\":\n",
    "    JOIN_STATISTICS_PATH = f\"~/Dokumente/repos/example_plugin/stats/final/tpch/sf10/01-l_orderkey-100/joins.csv\"\n",
    "elif BENCHMARK == \"TPC-DS\":\n",
    "    JOIN_STATISTICS_PATH = f\"~/Dokumente/repos/example_plugin/stats/final/tpcds/sf10-2d/nosort/joins.csv\"\n",
    "else:\n",
    "    raise Exception(\"Unknown benchmark: \" + BENCHMARK)\n",
    "\n",
    "# TODO: when using more than one run: multiply operator runtimes with their query frequencies\n",
    "\n",
    "ALL = 0\n",
    "ONLY_CLUSTERING = 1\n",
    "NO_CLUSTERING = 2\n",
    "CLUSTERING_STATE = 0\n",
    "\n",
    "joins = pd.read_csv(JOIN_STATISTICS_PATH, sep='|')\n",
    "joins = joins.dropna()\n",
    "if CLUSTERING_STATE == ONLY_CLUSTERING:\n",
    "    joins = joins[joins['RADIX_BITS'] > 0]\n",
    "elif CLUSTERING_STATE == NO_CLUSTERING:\n",
    "    joins = joins[joins['RADIX_BITS'] == 0]\n",
    "\n",
    "sum_columns = ['BUILD_SIDE_MATERIALIZING_NS', 'PROBE_SIDE_MATERIALIZING_NS', 'CLUSTERING_NS', 'BUILDING_NS', 'PROBING_NS', 'OUTPUT_WRITING_NS']\n",
    "sums = []\n",
    "sum_names = list(map(lambda x: x[:-3].replace('_', ' '), sum_columns))\n",
    "\n",
    "for sum_column in sum_columns:\n",
    "    sums.append(joins[sum_column].sum() / 1e6)\n",
    "\n",
    "\n",
    "#fig, ax = plt.subplots()\n",
    "fig = plt.figure()\n",
    "ax = fig.add_axes([0, 0, 1, 1])\n",
    "ax.bar(sum_names, sums)\n",
    "ax.get_yaxis().set_major_formatter(matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x), ',')))\n",
    "ax.set_xlabel('Hash join step')\n",
    "ax.set_ylabel('Run time in milliseconds')\n",
    "ax.set_title(f'Accumulated run times per hash join step over all {BENCHMARK} queries')\n",
    "plt.xticks(rotation=90)\n",
    "\n",
    "\n",
    "filename = '/home/aloeser/Downloads/hash_join_steps_time'\n",
    "if CLUSTERING_STATE == ALL:\n",
    "    filename += \"_all\"\n",
    "elif CLUSTERING_STATE == ONLY_CLUSTERING:\n",
    "    filename += \"_only_clustering\"\n",
    "elif CLUSTERING_STATE == NO_CLUSTERING:\n",
    "    filename += \"_no_clustering\"\n",
    "else:\n",
    "    assert False\n",
    "    \n",
    "filename += f\"_{BENCHMARK.lower().replace('-', '')}\"\n",
    "plt.savefig(f\"{filename}.pdf\", bbox_inches=\"tight\")\n",
    "\n",
    "\n",
    "print(sum(sums))\n",
    "print(joins['RUNTIME_NS'].sum() / 1e6)\n",
    "print(len(joins))\n",
    "\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_average_latencies(benchmark_folder):\n",
    "    latencies = []\n",
    "    for file in sorted(glob.glob(f\"{benchmark_folder}/*.json\")):\n",
    "        with open(file, 'r') as f:\n",
    "            benchmark_json = json.load(f)\n",
    "        latencies.append(extract_average_latency(benchmark_json))\n",
    "    \n",
    "    return latencies\n",
    "\n",
    "    \n",
    "def extract_average_latency(benchmark_json, num_runs=20):\n",
    "    benchmarks = benchmark_json['benchmarks']\n",
    "    total_latency = 0\n",
    "    \n",
    "    for benchmark in benchmarks:\n",
    "        assert len(benchmark['successful_runs']) == num_runs, f\"Expected {num_runs} succesful runs per query\"\n",
    "        for run in benchmark['successful_runs']:\n",
    "            total_latency += float(run['duration'])\n",
    "            \n",
    "    return total_latency / num_runs\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_cluster_counts(benchmark_folder, clustering_columns):\n",
    "    all_counts = []\n",
    "    for file in sorted(glob.glob(f\"{benchmark_folder}/*.json\")):\n",
    "        with open(file, 'r') as f:\n",
    "            benchmark_json = json.load(f)\n",
    "        config = benchmark_json['clustering_info']['config']\n",
    "        if \"tpch\" in benchmark_folder:\n",
    "            config = config['lineitem']\n",
    "        elif \"tpcds\" in benchmark_folder:\n",
    "            config = config['store_sales']\n",
    "        else:\n",
    "            raise Exception(\"unknown benchmark\")        \n",
    "        \n",
    "        counts = [1] * len(clustering_columns)\n",
    "        for entry in config:\n",
    "            # Skip sort column\n",
    "            if entry[1] > 1:\n",
    "                counts[clustering_columns.index(entry[0])] = entry[1]\n",
    "        all_counts.append(counts)\n",
    "    return pd.DataFrame(all_counts, columns=clustering_columns)\n",
    "            \n",
    "    \n",
    "\n",
    "#extract_cluster_counts(\"/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpch/sf10-2d\", ['l_receiptdate', 'l_shipdate', 'l_quantity', 'l_discount', 'l_partkey', 'l_suppkey', 'l_orderkey'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_cluster_count_table(benchmark_folder, potential_clustering_columns, column_names, ax):\n",
    "    def format_cluster_count(row):\n",
    "        num_clusters = row['num_clusters']\n",
    "        if 1.8 *  num_clusters < default_chunks:\n",
    "            new_cluster_count = np.int64(row[sort_column] * round(default_chunks / row['num_clusters']))\n",
    "            return str(new_cluster_count)\n",
    "            #return f\"{row[sort_column]} ({new_cluster_count})\"\n",
    "        else:\n",
    "            return str(row[sort_column])\n",
    "\n",
    "    table_df = extract_cluster_counts(benchmark_folder, potential_clustering_columns)\n",
    "    table_df['num_clusters'] = table_df.product(axis=1)\n",
    "    #table_df[sort_column] = table_df.apply(format_cluster_count, axis=1)\n",
    "    table_df.drop('num_clusters', inplace=True, axis=1)\n",
    "    table_df = table_df.transpose()\n",
    "\n",
    "    the_table = ax.table(cellText=table_df.values,\n",
    "                          colLabels=column_names,\n",
    "                          rowLabels=potential_clustering_columns,\n",
    "                          loc='center', bbox=[0, -0.0842 * len(potential_clustering_columns), 1, len(potential_clustering_columns) * 0.0821]) #\n",
    "\n",
    "\n",
    "    ax.set_xticks([])\n",
    "    for index, column in enumerate(potential_clustering_columns):\n",
    "        for rank in range(len(column_names)):        \n",
    "            if pd.isna(table_df.loc[column, rank]) or table_df.loc[column, rank] == 1:\n",
    "                the_table[ (index+1, rank)].set_text_props(text=\"\")\n",
    "            #else:\n",
    "            #    the_table[ (index+1, rank)].get_text().set_weight('bold')\n",
    "    return the_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[113365.45704265 113365.45704265 135667.33355065 121118.00137395\n",
      " 121015.11276085 122106.47233335 137730.7366561  123642.12992575\n",
      " 112489.21380875 115618.5784726  116044.3052439  120957.1315512\n",
      " 134884.93365095 133993.6602361  122708.91399375 124167.96190715\n",
      " 116078.3695687  115980.3996175  120665.42383085 134894.6425131\n",
      " 136919.44096115]\n",
      "[113365.45704265 113248.24678595 135667.33355065 124036.04273965\n",
      " 123894.1249488  121044.02141945 122415.1900698  120724.49817775\n",
      " 116775.5972383  125409.47944215 125078.3983255  121015.11276085\n",
      " 122327.47957445 121114.83936505 122392.21302405 124901.2653812\n",
      " 121764.28139485 137730.7366561  126297.1531181  119824.00074085\n",
      " 123642.12992575]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#latencies =      [112572, 134109, 119582, 122008, 123642, 111577, 114712, 119942, 115359, 120596, 121851, 123653, 116027, 120127, 115513, 134484, 139638, 132194, 134679, 131486, 136919]\n",
    "BENCHMARK = \"tpch\"\n",
    "THREE_D = True\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,5))\n",
    "if BENCHMARK == \"tpch\":\n",
    "    #latencies_2d = [113365, 113365, 135667, 121118, 122106, 123642, 122489, 115618, 120957, 116044, 121015, 122708, 124168, 116078, 120665, 115980, 134894, 140735, 132876, 133818, 132028]\n",
    "    #latencies_3d = [113248, 113248, 135123, 124036, 123894, 120651, 122415, 121044, 120724, 116775, 122392, 122956, 124901, 121764, 126297, 119824, 123250, 123614, 124119, 121595, 120524]\n",
    "    latency_shuffled = 144684\n",
    "    ax.set_ylim(ymin=100000,ymax=150000)\n",
    "    benchmark_folder = \"/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpch/sf10-2d\"\n",
    "    benchmark_folder_3d = \"/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpch/sf10-3d-corrected\"\n",
    "    potential_clustering_columns = ['l_receiptdate', 'l_shipdate', 'l_quantity', 'l_discount', 'l_partkey', 'l_suppkey', 'l_orderkey']\n",
    "    sort_column = \"l_orderkey\"\n",
    "    default_chunks = 916\n",
    "elif BENCHMARK == \"tpcds\":\n",
    "    #latencies_2d = [83172, 87656, 83363, 80135, 82968, 83763, 83211, 83349, 83822, 90347, 83299, 83155, 88603, 84389, 88911, 87934, 89292, 89534, 89033, 90669, 88279]\n",
    "    #latencies_3d = []\n",
    "    latency_shuffled = 94393\n",
    "    ax.set_ylim(ymin=70000,ymax=100000)\n",
    "    benchmark_folder = \"/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpcds/sf10-2d\"\n",
    "    benchmark_folder_3d = \"/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpcds/sf10-3d-corrected\"\n",
    "    potential_clustering_columns = ['ss_sold_date_sk', 'ss_item_sk', 'ss_hdemo_sk', 'ss_store_sk', 'ss_customer_sk', 'ss_addr_sk', 'ss_cdemo_sk', 'ss_promo_sk', 'ss_sold_time_sk', 'ss_ticket_number']\n",
    "    sort_column = \"ss_ticket_number\"\n",
    "    default_chunks = 440\n",
    "else:\n",
    "    raise Exception(\"unknown benchmark\")\n",
    "\n",
    "MILLISECONDS = 1e6\n",
    "latencies = np.array(extract_average_latencies(benchmark_folder)) / MILLISECONDS\n",
    "print(latencies)\n",
    "if THREE_D:\n",
    "    latencies_3d = np.array(extract_average_latencies(benchmark_folder_3d)) / MILLISECONDS\n",
    "    print(latencies_3d)\n",
    "\n",
    "ranks = ['Base'] + [str(x + 1) for x in range(20)]\n",
    "\n",
    "\n",
    "ax.plot(latencies, label='Measured latency 2d clusterings', linestyle='dotted', marker='.')\n",
    "if THREE_D:\n",
    "    ax.plot(latencies_3d, label='Measured latency 3d clusterings', linestyle='dotted', marker='.')\n",
    "\n",
    "\n",
    "plt.hlines(latency_shuffled, 0, 20, color='r', label='Latency on shuffled data', linestyle='dotted')\n",
    "min_latency = min(latencies)\n",
    "if THREE_D:\n",
    "    min_latency = min(min_latency, min(latencies_3d))\n",
    "plt.hlines(min(latencies), 0, 20, color='gray', label='Latency of the best suggested clustering', linestyle='dotted')\n",
    "if not THREE_D and BENCHMARK == \"tpch\" and False:\n",
    "     plt.hlines(107146, 0, 20, color='green', label='Latency of the best known clustering', linestyle='dotted')\n",
    "\n",
    "ax.get_yaxis().set_major_formatter(matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x), ',')))\n",
    "ax.set_ylabel('Latency (ms)')\n",
    "ax.set_xticks([])\n",
    "#ax.set_xticklabels(ranks)\n",
    "#ax.set_xlabel(\"Suggestion rank\")\n",
    "\n",
    "PAPER_MODE = True\n",
    "if PAPER_MODE:\n",
    "    box = ax.get_position()\n",
    "    ax.set_position([box.x0, box.y0, box.width, box.height * 0.8])\n",
    "    legend = ax.legend(loc=\"lower center\", ncol=2, bbox_to_anchor=(0.5, 1))\n",
    "else:\n",
    "    legend = ax.legend(loc='center left', ncol=1, bbox_to_anchor=(1.0, 0.5))\n",
    "\n",
    "if THREE_D:\n",
    "    the_table = add_cluster_count_table(benchmark_folder_3d, potential_clustering_columns, ranks, ax)\n",
    "else:\n",
    "    the_table = add_cluster_count_table(benchmark_folder, potential_clustering_columns, ranks, ax)\n",
    "\n",
    "plt.margins(1/42, 1/42) # no empty space on left and right within bar chart\n",
    "\n",
    "dimension_text = \"_3d\" if THREE_D else \"\"\n",
    "plt.savefig(f'/home/aloeser/Downloads/top20_latency_{BENCHMARK}{dimension_text}.pdf', bbox_extra_artists=(legend,the_table,), bbox_inches=\"tight\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[113365.45704265 115242.0721039  116085.11395195 111104.5338641\n",
      " 109309.9847971  107439.94468245 107146.13394665]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "benchmark_folder = \"/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpch/sf10-clustercounts\"\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "latencies = np.array(extract_average_latencies(benchmark_folder)) / 1e6\n",
    "print(latencies)\n",
    "ax.plot(np.arange(len(latencies)), latencies, linestyle='dotted', marker='.', label='Measured latency')\n",
    "ax.hlines(latencies[0], 0, 6, color='gray', label='Latency of the base clustering', linestyle='dotted')\n",
    "\n",
    "\n",
    "ranks = ['Base'] + [str(x + 1) for x in range(len(latencies)-1)]\n",
    "the_table = add_cluster_count_table(benchmark_folder, ['l_shipdate', 'l_orderkey'], ranks, ax)\n",
    "\n",
    "ax.get_yaxis().set_major_formatter(matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x), ',')))\n",
    "ax.set_ylabel('Latency (ms)')\n",
    "ax.set_ylim(ymin=100000,ymax=120000)\n",
    "\n",
    "PAPER_MODE = True\n",
    "if PAPER_MODE:\n",
    "    box = ax.get_position()\n",
    "    ax.set_position([box.x0, box.y0, box.width, box.height * 0.8])\n",
    "    legend = ax.legend(loc=\"lower center\", ncol=2, bbox_to_anchor=(0.5, 1))\n",
    "else:\n",
    "    legend = ax.legend(loc='center left', ncol=1, bbox_to_anchor=(1.0, 0.5))\n",
    "    \n",
    "plt.margins(1/15, 1/15)\n",
    "plt.savefig(f'/home/aloeser/Downloads/eval_cluster_counts.pdf', bbox_extra_artists=(legend,the_table,), bbox_inches=\"tight\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_operator_durations_from_pqp(pqp_folder):\n",
    "    all_operator_breakdowns = {}\n",
    "    for file in sorted(glob.glob(f\"{pqp_folder}/*-PQP.svg\")):\n",
    "        operator_breakdown = {}\n",
    "        with open(file, \"r\") as svg:\n",
    "            svg_string = svg.read().replace(\"\\n\", \"|\")\n",
    "\n",
    "            # Find the \"total by operator\" table using a non-greedy search until the end of the <g> object\n",
    "            table_string = re.findall(r\"Total by operator(.*?)</g>\", svg_string)[0]\n",
    "\n",
    "            # Replace all objects within the table string, also trim newlines (rewritten to |) at the begin and the end\n",
    "            table_string = re.sub(r\"<.*?>\", \"\", table_string)\n",
    "            table_string = re.sub(r\"^\\|*\", \"\", table_string)\n",
    "            table_string = re.sub(r\"\\|*$\", \"\", table_string)\n",
    "\n",
    "            row_strings = table_string.split(\"||\")\n",
    "\n",
    "            # The svg table stores data in a columnar orientation, so we first extract the operator names, then their\n",
    "            # durations\n",
    "            operator_names = []\n",
    "            operator_durations = []\n",
    "\n",
    "            for operator_name in row_strings[0].split(\"|\"):\n",
    "                operator_names.append(operator_name.strip())\n",
    "\n",
    "            # Convert time string to nanoseconds and add to operator_durations\n",
    "            for operator_duration_str in row_strings[1].split(\"|\"):\n",
    "                operator_duration = pd.Timedelta(operator_duration_str.replace(\"µ\", \"u\")).total_seconds() * 1e9\n",
    "                operator_durations.append(operator_duration)\n",
    "\n",
    "            operator_breakdown = dict(zip(operator_names, operator_durations))\n",
    "\n",
    "            # Ignore the \"total\" line\n",
    "            del operator_breakdown[\"total\"]\n",
    "\n",
    "        # Store in all_operator_breakdowns\n",
    "        all_operator_breakdowns[file.replace(\"-PQP.svg\", \"\").replace(\"TPC-H_\", \"Q\")] = operator_breakdown\n",
    "\n",
    "    # Make operators the columns and order by operator name\n",
    "    df = pd.DataFrame(all_operator_breakdowns).transpose()\n",
    "    df = df.reindex(sorted(df.columns, reverse=True), axis=1)\n",
    "\n",
    "    df = df.fillna(0)\n",
    "\n",
    "    # Calculate share of total execution time (i.e., longer running benchmark items are weighted more)\n",
    "    #df.loc[\"Absolute\"] = df.sum() / df.count()\n",
    "\n",
    "    # Normalize data from nanoseconds to milliseconds\n",
    "    df.iloc[:, 0:] = df.iloc[:, 0:].apply(lambda x: x / 1000000, axis=1)\n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_operator_durations_from_csv(csv_folder):\n",
    "    scan_df = pd.read_csv(csv_folder + \"/table_scans.csv\", sep='|')\n",
    "    if \"tpch\" in csv_folder:\n",
    "        scan_df = scan_df[scan_df['TABLE_NAME'] == \"lineitem\"]\n",
    "    elif \"tpcds\" in csv_folder:\n",
    "        scan_df = scan_df[scan_df['TABLE_NAME'] == \"store_sales\"]\n",
    "    scan = scan_df['RUNTIME_NS'].sum()\n",
    "    \n",
    "    \n",
    "    aggregate_columns = \"OPERATOR_TYPE|QUERY_HASH|OPERATOR_HASH|LEFT_INPUT_OPERATOR_HASH|RIGHT_INPUT_OPERATOR_HASH|COLUMN_TYPE|TABLE_NAME|COLUMN_NAME|GROUP_BY_COLUMN_COUNT|AGGREGATE_COLUMN_COUNT|INPUT_CHUNK_COUNT|INPUT_ROW_COUNT|OUTPUT_CHUNK_COUNT|OUTPUT_ROW_COUNT|GROUP_BY_KEY_PARTITIONING_NS|AGGREGATING_NS|GROUP_BY_COLUMNS_WRITING_NS|AGGREGATE_COLUMNS_WRITING_NS|OUTPUT_WRITING_NS|RUNTIME_NS|DESCRIPTION\".split(\"|\")\n",
    "    aggregate_df = pd.read_csv(csv_folder + \"/aggregates.csv\", sep='|', names=aggregate_columns)\n",
    "    aggregate_operators = aggregate_df.groupby(['QUERY_HASH', 'OPERATOR_HASH'])['RUNTIME_NS'].mean()\n",
    "    aggregate = aggregate_operators.sum()\n",
    "    \n",
    "    join_df = pd.read_csv(csv_folder + \"/joins.csv\", sep='|')\n",
    "    join_estimated = (join_df['BUILD_SIDE_MATERIALIZING_NS'] + join_df['PROBE_SIDE_MATERIALIZING_NS'] + join_df['PROBING_NS']).sum()\n",
    "    join_unestimated = join_df['RUNTIME_NS'].sum() - join_estimated\n",
    "    \n",
    "    return scan, join_estimated, join_unestimated, aggregate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_operator_durations_range(folder):\n",
    "    aggregates = []\n",
    "    joins_estimated = []\n",
    "    joins_unestimated = []\n",
    "    scans = []\n",
    "    for csv_folder in sorted(glob.glob(f\"{folder}/*\")):\n",
    "        #print(f\"processing {csv_folder}\")\n",
    "        runtimes = extract_operator_durations_from_csv(csv_folder)\n",
    "        aggregates.append(int(runtimes[3]))\n",
    "        joins_estimated.append(int(runtimes[1]))\n",
    "        joins_unestimated.append(int(runtimes[2]))\n",
    "        scans.append(int(runtimes[0]))\n",
    "    \n",
    "    return np.array(aggregates), np.array(joins_estimated), np.array(joins_unestimated), np.array(scans)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "BENCHMARK = \"tpch\"\n",
    "if BENCHMARK == \"tpch\":\n",
    "    #guesses = [47420, 47420, 47998, 48936, 48952, 48960, 48963, 48965, 48965, 48965, 49008, 49020, 49028, 49037, 49037, 49037, 49371, 49376, 49400, 49400, 49400]\n",
    "    guesses = [47947, 47947, 49211, 49569, 49640, 49897, 49901, 49905, 49909, 49911, 49911, 49911, 49934, 49945, 49964, 49973, 49982, 49982, 49982, 50011, 50041]\n",
    "    stats_path = '/home/aloeser/Dokumente/repos/example_plugin/stats/final/tpch/sf10-2d'\n",
    "    benchmark_folder = '/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpch/sf10-2d'\n",
    "    potential_clustering_columns = ['l_receiptdate', 'l_shipdate', 'l_quantity', 'l_discount', 'l_partkey', 'l_suppkey', 'l_orderkey']\n",
    "    table_name = \"lineitem\"\n",
    "elif BENCHMARK == \"tpcds\":    \n",
    "    guesses = [33728, 33628, 33728, 33743, 33743, 33772, 33772, 33778, 33787, 33787, 33787, 33787, 33939, 33939, 33968, 33968, 33978, 33983, 33983, 33983, 33983]    \n",
    "    stats_path = '/home/aloeser/Dokumente/repos/example_plugin/stats/final/tpcds/sf10-2d-correction'\n",
    "    benchmark_folder = '/home/aloeser/Dokumente/repos/hyrise/benchmark_results/final/tpcds/sf10-2d'\n",
    "    potential_clustering_columns = ['ss_sold_date_sk', 'ss_item_sk', 'ss_hdemo_sk', 'ss_store_sk', 'ss_customer_sk', 'ss_addr_sk', 'ss_cdemo_sk', 'ss_promo_sk', 'ss_sold_time_sk', 'ss_ticket_number']\n",
    "    table_name = \"store_sales\"\n",
    "else:\n",
    "    raise Exception(\"unknown benchmark\")\n",
    "\n",
    "    \n",
    "aggregates, joins_estimated, joins_unestimated, scans = extract_operator_durations_range(stats_path)\n",
    "MS_FACTOR = 1e6\n",
    "aggregates = aggregates / MS_FACTOR\n",
    "joins_estimated = joins_estimated / MS_FACTOR\n",
    "joins_unestimated = joins_unestimated / MS_FACTOR\n",
    "scans = scans / MS_FACTOR        \n",
    "    \n",
    "#scans[0] = scans[1]\n",
    "#joins_estimated[0] = joins_estimated[1]\n",
    "#joins_unestimated[0] = joins_unestimated[1]\n",
    "#aggregates[0] = aggregates[1]\n",
    "\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,5))\n",
    "ax.get_yaxis().set_major_formatter(matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x), ',')))\n",
    "x = np.arange(len(scans))\n",
    "xlabels = ['Base'] + [str(x) for x in range(1,21)]\n",
    "scanbar = ax.bar(x, scans, label=f'Measured latency of scans on {table_name}')\n",
    "join_estimated_bar = ax.bar(x, joins_estimated, bottom=scans, label='Measured latency of estimated join steps')\n",
    "join_unestimated_bar = ax.bar(x, joins_unestimated, bottom=joins_estimated+scans, label='Measured latency of join steps unaffected by the clustering')\n",
    "aggregatebar = ax.bar(x, aggregates, bottom=joins_unestimated+joins_estimated+scans, label='Measured latency of aggregates')\n",
    "ax.plot(guesses, color=\"black\", linestyle='dotted',marker='.', label='Estimated latency of scans and joins')\n",
    "#ax.set_xticks(x)\n",
    "ax.set_xticks([])\n",
    "#ax.set_xticklabels(xlabels)\n",
    "#ax.set_xlabel('Suggestion rank')\n",
    "ax.set_ylabel('Latency (ms)')\n",
    "ranks = ['Base'] + [str(x + 1) for x in range(20)]\n",
    "\n",
    "PAPER_MODE = True\n",
    "if PAPER_MODE:\n",
    "    box = ax.get_position()\n",
    "    ax.set_position([box.x0, box.y0, box.width, box.height * 0.8])\n",
    "    legend = ax.legend(loc=\"lower center\", ncol=1, bbox_to_anchor=(0.5, 1))\n",
    "else:\n",
    "    legend = ax.legend(loc='center left', ncol=1, bbox_to_anchor=(1.0, 0.5))\n",
    "\n",
    "\n",
    "the_table = add_cluster_count_table(benchmark_folder, potential_clustering_columns, ranks, ax)\n",
    "\n",
    "plt.margins(0.005, 0.005) # no empty space on left and right within bar chart\n",
    "plt.savefig(f'/home/aloeser/Downloads/detailed_estimates_{BENCHMARK}.pdf', bbox_extra_artists=(legend,the_table), bbox_inches=\"tight\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0052716827953998\n",
      "45721.035579 55783.95934100001\n",
      "[37214.485801 37214.485801 49460.265298 46911.246311 45435.409247\n",
      " 46804.415724 50036.753859 45534.143942 42095.956827 43503.273345\n",
      " 43917.928125 46961.800039 49944.45753  48501.571101 46494.319481\n",
      " 45333.965722 43967.565136 44077.80756  45917.33627  47812.057651\n",
      " 49205.248493]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([0.03101351, 0.03101351, 0.02999022, 0.00873278, 0.00911851,\n",
       "       0.00918987, 0.02680925, 0.00603289, 0.01121294, 0.00697966,\n",
       "       0.00862887, 0.00830961, 0.03015774, 0.02652015, 0.00563422,\n",
       "       0.00603297, 0.00749405, 0.00915756, 0.00819484, 0.02599984,\n",
       "       0.02553504])"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "measured_latency = scans + joins_estimated + joins_unestimated\n",
    "\n",
    "print(stats.gmean(measured_latency / guesses))\n",
    "print(min(measured_latency), max(measured_latency))\n",
    "print(aggregates)\n",
    "scans/measured_latency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "BENCHMARK = \"tpcds\"\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "\n",
    "column_to_counts = {}\n",
    "if BENCHMARK == \"tpch\":\n",
    "    column_to_counts['l_shipdate'] = np.array([3,4,5,6,6,7,8,9,10,11,12,14,14,16,16,17,20,20,20,21,24,24,24,24,24,26,32,32,32,32,32,32,32,32,32,34,36,47,48,48,48,48,48,48,48,48,48,48,48,48,48,48,48,48,48,48,48,48,49,49,58,59,61,94,94,94,94,94,94,94,94,94,94,95,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,96,])\n",
    "    column_to_counts['l_orderkey'] = np.array([2,4,4,5,6,8,8,10,10,12,13,13,15,15,17,17,17,20,20,20,20,25,25,25,25,25,25,34,34,34,34,34,34,34,34,34,34,34,42,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,])\n",
    "elif BENCHMARK == \"tpcds\":\n",
    "    column_to_counts['ss_ticket_number'] = np.array([2,4,4,5,6,8,8,10,10,12,13,13,15,15,17,17,17,20,20,20,20,25,25,25,25,25,25,34,34,34,34,34,34,34,34,34,34,34,41,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100,])\n",
    "    column_to_counts['ss_sold_time_sk'] = np.array([4,4,6,6,8,9,9,10,12,12,13,15,15,17,17,18,19,21,21,22,24,24,24,25,28,29,31,32,32,33,33,34,34,35,36,36,36,40,41,41,41,49,51,53,53,53,53,54,54,54,54,56,56,58,58,58,61,63,63,63,63,64,68,77,77,77,77,77,77,77,78,78,81,81,81,83,85,85,85,85,85,85,85,86,86,86,86,86,86,86,86,86,86,86,86,86,86,86,86,])\n",
    "else:\n",
    "    raise Exception(\"unknown benchmark\")\n",
    "\n",
    "ideal_boundaries = np.arange(2,101)\n",
    "    \n",
    "NORMALIZE = False\n",
    "if NORMALIZE:    \n",
    "    tmp = {}\n",
    "    for column, counts in column_to_counts.items():\n",
    "        tmp[column] = counts / ideal_boundaries\n",
    "    column_to_counts = tmp\n",
    "    \n",
    "    ideal_boundaries = ideal_boundaries / ideal_boundaries\n",
    "    ax.set_ylabel('Relative error')    \n",
    "else:\n",
    "    ax.set_ylabel(\"Obtained boundary count\")\n",
    "    \n",
    "\n",
    "for column, counts in column_to_counts.items():\n",
    "    ax.plot(counts, label=column, linestyle='dotted', marker='.')\n",
    "ax.plot(ideal_boundaries, label='Ideal number of boundaries', linestyle='dashed')\n",
    "ax.set_xlabel('Desired boundary count')\n",
    "\n",
    "\n",
    "PAPER_MODE = True\n",
    "if PAPER_MODE:\n",
    "    box = ax.get_position()\n",
    "    ax.set_position([box.x0, box.y0, box.width, box.height * 0.8])\n",
    "    legend = ax.legend(loc=\"lower center\", ncol=2, bbox_to_anchor=(0.5, 1))\n",
    "else:\n",
    "    legend = ax.legend(loc='center left', ncol=1, bbox_to_anchor=(1.0, 0.5))\n",
    "\n",
    "extension = '_normalized' if NORMALIZE else ''\n",
    "plt.savefig(f'/home/aloeser/Downloads/cluster_boundaries_counts_{BENCHMARK}{extension}.pdf', bbox_inches='tight', bbox_extra_artists=(legend,))\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "shipdate100          = [78, 67, 58, 48, 44, 38, 31, 25, 21, 20, ]\n",
    "shipdate10orderkey30 = [42,  2,  0, 0, 0, 0, 0, 0, 0, 0, ] \n",
    "shipdate10orderkey20 = [53, 14, 5, 1, 1, 1, 0, 0, 0, 0, ]\n",
    "shipdate10orderkey15 = [33, 16, 9, 4, 3, 2, 0, 0, 0, 0, ]\n",
    "\n",
    "\n",
    "\n",
    "INCLUDE_START = True\n",
    "if INCLUDE_START:\n",
    "    shipdate100 = [96] + shipdate100\n",
    "    shipdate10orderkey30 = [300] + shipdate10orderkey30\n",
    "    shipdate10orderkey20 = [200] + shipdate10orderkey20\n",
    "    shipdate10orderkey15 = [150] + shipdate10orderkey15\n",
    "\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "\n",
    "x = np.arange(len(shipdate100))\n",
    "ax.set_xlabel('Attempt')\n",
    "ax.set_xticks(x)\n",
    "x_ticklabels = [str(x) for x in x]\n",
    "ax.set_xticklabels(x_ticklabels)\n",
    "\n",
    "ax.set_ylabel('Unsorted clusters')\n",
    "\n",
    "ax.plot(shipdate100, label='9 chunks per cluster', linestyle='dotted', marker='.')\n",
    "ax.plot(shipdate10orderkey15, label='6 chunks per cluster', linestyle='dotted', marker='.')\n",
    "ax.plot(shipdate10orderkey20, label='4.5 chunks per cluster', linestyle='dotted', marker='.')\n",
    "ax.plot(shipdate10orderkey30, label='3 chunks per cluster', linestyle='dotted', marker='.')\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "PAPER_MODE = True\n",
    "if PAPER_MODE:\n",
    "    box = ax.get_position()\n",
    "    ax.set_position([box.x0, box.y0, box.width, box.height * 0.8])\n",
    "    legend = ax.legend(loc=\"lower center\", ncol=2, bbox_to_anchor=(0.5, 1))\n",
    "else:\n",
    "    legend = ax.legend(loc='center left', ncol=1, bbox_to_anchor=(1.0, 0.5))\n",
    "\n",
    "plt.savefig(f'/home/aloeser/Downloads/failed_steps_after_attempt.pdf', bbox_inches='tight', bbox_extra_artists=(legend,))\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "CONFIG = 1\n",
    "SF = 10\n",
    "if SF == 1:    \n",
    "    bytes_used = [563591860, 2800991212, 2825378783, 2853603645, 2874713188, 2902250511, 2930221195, 2959492291, 3198286082, 3961221241, 4737458079, 5193840845, 4864270448, 4540017522, 4220256432, 3906121512, 3579064980, 580077332, 580077332, ]\n",
    "    times = [37779863, 5093477753, 10152086955, 15228348282, 20335170980, 25442738584, 30571112339, 35727253005, 40912784847, 46182718470, 51461759775, 56764337432, 62048378019, 67308542176, 72545971671, 77761596151, 82956194119, 88015251180, 93065931427, ]\n",
    "    phase_ends = [39486068750, 39486167232, 54961857667, 83323097011]\n",
    "\n",
    "    bytes_used_delete = [563591860, 2800147389, 2764496406, 2724525991, 2670664041, 2629821659, 2595474154, 2542515660, 2537878180, 3301116374, 3267559680, 2362652979, 1987040259, 1664878439, 1352187921, 1043921426, 731110511, 580077332, 580077332, ]\n",
    "    times_delete = [38152849, 5093386854, 10149759694, 15222055279, 20318959493, 25411908687, 30524322368, 35664264162, 40823891085, 46032811330, 51236046684, 56417380210, 61548048640, 66660123349, 71750956831, 76820661608, 81869715566, 86917755431, 91967644113, ]\n",
    "    phase_ends_delete = [40524493366, 40524572571, 55207957598, 84337879953]\n",
    "elif SF == 10:\n",
    "    lineitem_dictionary_bytes = 5101742961\n",
    "    lineitem_value_bytes = 23293954731\n",
    "    if CONFIG == 1:    \n",
    "        bytes_used = [6163033371, 7353824544, 7378907660, 7424845565, 7474925735, 9642471301, 9734609092, 9825662067, 9871273380, 9917381327, 12084036538, 12176645287, 12221749874, 12289164072, 12843626577, 14527396068, 14621228648, 14667180912, 14714138780, 16860252290, 16999013580, 17070383065, 17119131648, 17673824688, 19360290535, 19431385533, 19525818756, 20057846828, 21744067151, 21793471153, 21912390100, 22098856705, 22540319499, 24205119287, 24255536739, 24375632005, 24932614963, 26596936785, 26671390092, 26745324148, 27326185657, 28992656663, 29066279052, 29109919327, 30540008493, 31795674098, 33055104076, 34313196435, 35793667839, 37001649022, 38453322583, 39662641482, 41114933289, 42565838094, 43773119120, 45223534834, 46674224317, 48126627893, 49576338253, 51025742024, 52247364710, 51562676705, 50886972151, 50216829929, 49576069803, 48915181148, 48248368948, 47587001430, 46938701388, 46285542749, 45712151700, 45065846426, 44424660871, 43774300615, 43137688176, 42496687249, 41859960854, 41222784970, 40587331407, 39955317266, 39315111172, 38686120135, 38061338720, 37445182477, 36826380427, 36202415113, 35590994780, 32641821419, 12421944877, 6028716036, ]\n",
    "        times = [323802022, 8664399895, 16983795112, 25329269694, 33715699809, 42144101015, 50595950565, 59080890457, 67589399526, 76137969474, 84726263715, 93337918504, 101969026739, 110631576672, 119339142558, 128088844310, 136861128739, 145655902810, 154487605884, 163370313561, 172278551198, 181210963647, 190179902157, 199190180361, 208232775299, 217298745455, 226396692410, 235536689536, 244716999050, 253916712638, 263141716355, 272417656015, 281733840389, 291083610454, 300467813311, 309885466800, 319327486430, 328807934769, 338331886966, 347889068846, 357487656086, 367111112900, 376762206101, 386459023330, 396240396925, 406098626246, 416036660121, 426049022781, 436153575485, 446330957533, 456955421727, 467295306697, 477727472809, 488243162518, 498833819226, 509513668648, 520283050656, 531140064890, 542086983370, 553119056024, 564248398103, 575306249704, 586317939653, 597283629097, 608208755053, 619086650958, 629915194837, 640698430421, 651438701675, 662133781433, 672927383651, 683535810903, 694101939065, 704623392270, 715103646888, 725536911033, 735928865968, 746273846504, 756602444757, 766861780409, 777080477298, 787252530204, 797384759676, 807475749264, 817524479171, 827528701707, 837492182844, 847443126494, 856150065194, 865042724862, ]\n",
    "        phase_ends = [384904808470, 384904948441, 562020818449, 844634279568]\n",
    "\n",
    "        bytes_used_delete = [6255319503, 7354093693, 7290824331, 7231282810, 7174502673, 9233721025, 9218090185, 9203881005, 9139001192, 9064236585, 11129663508, 11113720100, 11040548836, 10994085296, 11023062797, 13007569881, 12995653078, 12924828606, 12852797235, 13281008209, 14861407868, 14841274837, 14808847651, 14742697377, 15176474944, 16758010302, 16708255805, 16677400205, 17095408677, 18613948929, 18582460384, 18575361013, 18504393681, 18948110777, 20522139968, 20448108903, 20437544426, 20872135744, 21143393868, 22355874399, 22295818615, 22438624544, 22720642104, 24281065931, 24208973828, 24361567512, 24507086452, 24626437063, 24673387783, 24541476203, 25276382019, 25148249382, 25098826137, 25164868880, 25142245725, 24969942110, 24843032189, 24741501555, 24565702378, 24992449701, 24943665235, 24894215792, 24841221361, 24790766449, 24217434260, 22628969583, 22124331994, 21614052629, 21106615154, 20626355166, 20110072071, 19607418453, 19107720685, 18599416628, 18099485285, 17591086999, 17100330562, 16599100586, 16087832979, 15555108322, 15026460887, 14533808511, 14027613065, 13541398362, 13052828337, 12570854573, 12072561226, 11586920459, 11118930792, 10634667929, 10156840979, 9678486148, 9234501243, 8809807595, 8338347060, 7858178297, 7386493565, 6923861854, 6462045455, 6028716036, 6028716036, ]\n",
    "        times_delete = [338032805, 8700823937, 17020375195, 25366569763, 33745756149, 42164858966, 50605235277, 59061973106, 67550989403, 76153202491, 84707261223, 93284549761, 101887111651, 110613420680, 119274106709, 128075052054, 136788102456, 145522375226, 154288518564, 163207210935, 172169217971, 181153783524, 190159657074, 199195327889, 208125790944, 217102634266, 226090749091, 235110849731, 244161391336, 253254007773, 262370726737, 271647825775, 280950834812, 290305190710, 299545104603, 308935657037, 318225015400, 327546366212, 336914239008, 346305418298, 355835527249, 365401613002, 374877859586, 384391028617, 393917764075, 403485806409, 413062375237, 422643848077, 432229097965, 441996085518, 451621103437, 461238126385, 470854786132, 480478093730, 490183339186, 499800205669, 509569422403, 519298297036, 529149146608, 538873726208, 548588524460, 558341313448, 568070314141, 577797195151, 587574815503, 597113098169, 606577501223, 616003936934, 625397128834, 634761306715, 644090039105, 653383187150, 662638216637, 671861545067, 681051164163, 690205923007, 699329205138, 708414082413, 717467041550, 726483054559, 735461308785, 744403533051, 753311389268, 762194892417, 771039554661, 779849553618, 788626531294, 797367823349, 806078267095, 814754959111, 823400177839, 832012057486, 840684154341, 849235544166, 857755815588, 866245133443, 874701148189, 883126485608, 891521362852, 899885825180, 908358759785, ]\n",
    "        phase_ends_delete = [400535475826, 400535613422, 583680891236, 899699966405]\n",
    "        \n",
    "        bytes_used_delete_encoded = [6139961833, 7353891818, 7290471781, 7230962995, 7174238703, 9124016331, 8721298258, 8267334731, 7777019682, 7305240969, 7467582423, 8884824419, 8432360345, 7986883601, 7522743315, 7295760751, 7606899719, 8726556310, 8317684226, 7819296323, 7402372809, 7767064277, 8999361238, 8572556296, 8134585795, 7669505297, 7443064773, 7699255111, 8918760549, 8415283330, 8000639782, 7560049288, 7913029400, 7586984089, 8706544994, 8177778225, 7728332593, 7567345195, 7708266324, 8954775307, 8427696185, 7933138362, 7686907562, 7836894250, 7943284812, 8766857504, 8260984519, 7805318357, 8049748509, 7729608715, 9092906235, 8606147148, 8093738108, 8174948660, 8715895102, 8648152084, 9166453824, 9813758028, 10410673979, 10990372711, 11648279172, 12301056400, 12918957374, 13503570366, 14037694182, 14710081093, 15463710379, 16045311608, 16796988133, 16565366142, 15262153070, 14251364412, 13240481885, 12284298637, 11619116911, 11081227727, 10557365161, 10025870089, 9507576637, 8989236871, 8469718018, 7948752278, 7440094112, 6923967198, 6414024286, 6028716036, ]\n",
    "        times_delete_encoded = [320287606, 8663154553, 16980249386, 25322210994, 33702069877, 42108603004, 50591070509, 58976487582, 67358708863, 75789533518, 84197238523, 92602548796, 101053682723, 109443493172, 117883741635, 126281307832, 134756574108, 143218965154, 151671904880, 160113447731, 168553976492, 176974312978, 185419037657, 193878056485, 202326976872, 210771604677, 219231128075, 227713393678, 236135003601, 244540817272, 252938094602, 261387778970, 269808302026, 278289831042, 286711197488, 295112324165, 303563644465, 311989193908, 320404567032, 328830531640, 337256266038, 345720851105, 354203886984, 362622329625, 371130097967, 379565707027, 387981507964, 396389979216, 404884662094, 413377948617, 421827984747, 430320357060, 438753834294, 447267997446, 455827397476, 464418547652, 472989914292, 481603666077, 490253233732, 498948278604, 507687322967, 516469234412, 525297420294, 534158381100, 543050905216, 551990709129, 560977938190, 570009093054, 579082912377, 588146945887, 597127979213, 606041287625, 614887091627, 623666406148, 632400118569, 641098207827, 649762666760, 658389655040, 666983147589, 675541558628, 684063576524, 692552064115, 701004084177, 709421935614, 717806938721, 726217543869, ]\n",
    "        phase_ends_delete_encoded = [442253461003, 442253542793, 584102016743, 724043781666]\n",
    "\n",
    "    elif CONFIG == 2:\n",
    "        bytes_used = [7178180548, 28198509689, 28233432251, 28278058654, 28324420034, 28369861104, 28414636704, 28461348998, 28506373672, 28552005215, 28589861138, 28628816983, 28671130609, 28715632812, 28761287096, 28808209265, 28855590572, 28902268708, 28946367616, 28986644476, 29022731010, 29058682781, 29097553170, 29141480137, 29189982192, 29237735997, 29291571675, 29344754242, 29394520969, 29441315585, 29487237871, 29534300833, 29582385435, 29631263085, 29680207294, 29729889579, 29778715113, 29827654811, 29878134080, 29928044509, 29978671905, 30029481051, 30079792196, 30125801006, 30199625158, 31656720824, 33516480939, 35503887607, 37498591156, 39554076768, 41611471847, 43689701750, 45695860286, 47727596294, 49784354435, 51862404546, 53609839164, 52965425028, 52338370753, 51721778112, 51133386850, 50522314347, 49912621444, 49300817207, 48694646052, 48089752816, 47500940972, 46902501566, 46305296834, 45705366829, 45114507480, 44526172406, 43940629514, 43347598513, 42767385374, 42186656123, 41604143284, 41028357643, 40446856806, 39865580975, 39292317152, 38726367245, 38150633207, 37580562121, 37018238973, 36450553180, 5960901545, ]\n",
    "        times = [326223540, 8708293738, 17049060061, 25409500338, 33800576447, 42213701825, 50645805213, 59111851073, 67590866249, 76094796328, 84690397690, 93225165524, 101803330717, 110409695033, 119038872381, 127680966662, 136359497678, 145042740072, 153740969068, 162465277995, 171208859751, 180043206892, 188930577889, 197814421043, 206774037110, 215766220945, 224729014714, 233722606569, 242722936516, 251759183897, 260824318661, 269914801512, 279046815932, 288214816514, 297390697936, 306610967340, 315870064163, 325187224009, 334553028658, 344004677567, 353485224196, 362974580470, 372491396664, 382195681636, 391817610181, 401537392904, 411524019752, 421491586599, 431576507663, 441777640265, 452101020970, 462544664410, 473248041116, 484070527054, 495009518171, 506062270418, 517224735098, 528205229771, 539144370176, 550046602202, 560910898208, 571729770722, 582513514871, 593256925673, 603964212105, 614627329465, 625251943642, 635833488745, 646380045789, 656880361638, 667344020599, 677760630316, 688131728581, 698463506313, 708763132552, 719021495493, 729245500649, 739427979884, 749577564600, 759682317452, 769755304419, 779792669734, 789797742873, 799763604562, 809692263217, 819582591524, 828062392991, ]\n",
    "        phase_ends = [395194672921, 395204079892, 515729958748, 820564203126]\n",
    "\n",
    "        bytes_used_delete = [7132037490, 28198347607, 28161047595, 28065983065, 28040818448, 27985432833, 27919414173, 27869283342, 27808667393, 27740842476, 27687896493, 27609735869, 27549327036, 27489130705, 27420538427, 27347123656, 27290948635, 27231380034, 27174103652, 27103371324, 27027375624, 26955137792, 26872698869, 26806560193, 26732864713, 26670326969, 26600258324, 26542053826, 26480854336, 26422200389, 26351982261, 26275984107, 26167086166, 26134580800, 26062193683, 25988019612, 25930040795, 25866088405, 25790974126, 25734272247, 25659488146, 25585873941, 25509118519, 25450636822, 25376351700, 25304471360, 25315783189, 26422703826, 25914181837, 26613182253, 25900820892, 26506619394, 25297758459, 26552385512, 26832191029, 25596913319, 26778044903, 25112674906, 26646104189, 25524442667, 22499040043, 21959151380, 21419760404, 20895564447, 20394649271, 19873000500, 19348405888, 18821159250, 18300173411, 17778398199, 17296099550, 16842228228, 16362668083, 15884921676, 15402529665, 14925439727, 14458900745, 13998602562, 13518346316, 13045895556, 12577463154, 12105159986, 11634917324, 11163447053, 10692511646, 10222265872, 9755034230, 9288750228, 8828225950, 8369987084, 7902942177, 7449924221, 6984942910, 6532979162, 6073892119, 5960901545, ]\n",
    "        times_delete = [326678971, 8710204062, 17062729773, 25423294841, 33807866488, 42218994989, 50703659148, 59152300195, 67614125010, 76159822302, 84646968221, 93251196312, 101784468352, 110318757655, 118904302168, 127576766526, 136286864149, 145023659078, 153666049750, 162341945240, 171029649991, 179741193884, 188634448223, 197463882704, 206252628628, 215050868003, 224021364413, 232895537188, 241802703154, 250721024966, 259661169188, 268640409792, 277638192148, 286653481911, 295696370324, 304867553670, 313986859223, 323155866676, 332501623791, 341721181563, 351012044946, 360335276254, 369832215026, 379212009113, 388615663281, 398079367763, 407709464903, 417277871817, 426812288076, 436578463202, 446186097513, 455850014865, 465366406354, 474955262468, 484713690970, 494257121891, 503870734527, 513395469107, 523015022448, 532719695978, 542162629379, 551573822746, 560950694998, 570290841553, 579598930343, 588875608138, 598117715253, 607326323261, 616503567695, 625643338865, 634752158759, 643832747519, 652877589654, 661890621244, 670874230780, 679823608333, 688738892680, 697623165475, 706470768865, 715286268444, 724071244467, 732826266101, 741551276067, 750246158486, 758909212470, 767543261878, 776148923230, 784723026297, 793268553207, 801783446928, 810267800282, 818724271522, 827151055971, 835548942462, 843916608254, 852333614939, ]\n",
    "        phase_ends_delete = [408289226522, 408298643058, 528605000915, 845929969275]\n",
    "        \n",
    "        bytes_used_delete_encoded = [7155109010, 28198556457, 28161350356, 28089455951, 28043642910, 27998713875, 27931506263, 27880052893, 27818700615, 27759286402, 27702461273, 27625088833, 27562977467, 27498049576, 27426203274, 27364833159, 27311696342, 27239707199, 27184430969, 27114381861, 27035492572, 26970890674, 26895010926, 26814479020, 26751862173, 26656586781, 26620946787, 26549350081, 26498478960, 26433901966, 26367311138, 26261796948, 26212221853, 26150807774, 26079646210, 26009441994, 25938675415, 25876290800, 25803423402, 25739842343, 25669721434, 25592372693, 25521250798, 25457542628, 25383036936, 25311897917, 25331410501, 24122718763, 23898165628, 24185655244, 22816444031, 23181101004, 21581005104, 22187540453, 22313996187, 21803207651, 21952526095, 22106396922, 22347742851, 23285288206, 22204198605, 21164084015, 20323217455, 19423440954, 18542778393, 17762080971, 16953619862, 16095951024, 15330840763, 14483774523, 13656070472, 12815414449, 11975004833, 11172911246, 10393771088, 9582619854, 8763174662, 7935649329, 7131114749, 6387823897, 5960901545, ]\n",
    "        times_delete_encoded = [325161867, 8705929570, 17051326099, 25451416492, 33875474604, 42277181358, 50765072092, 59202571105, 67746281471, 76215509812, 84701775354, 93278405635, 101799971371, 110326229553, 118984449207, 127656199872, 136275017128, 145006284363, 153640177055, 162286660983, 171058514058, 179735556699, 188491368069, 197403684166, 206196804776, 214991492649, 223823069505, 232829683201, 241727040397, 250630165104, 259557875735, 268533864263, 277674381959, 286684138194, 295715684775, 304759435227, 313876191082, 323040237774, 332346736774, 341538988304, 350796436757, 360115927616, 369612874773, 378988891753, 388388793894, 397841221393, 407349535231, 417070543545, 426780370241, 436509744533, 446141852753, 455791078844, 465344341139, 474939329305, 484544808163, 493910428597, 503498857667, 513105760733, 522727099244, 532423112680, 542050418384, 551603342223, 560893977919, 570125777678, 579397755666, 588655801158, 597732435886, 606913580730, 616047373055, 625106682937, 633972316437, 642786642398, 651546026956, 660363254791, 669019522642, 677628738381, 686175519379, 694667389217, 703102509903, 711490851573, 719905245232, ]\n",
    "        phase_ends_delete_encoded = [408000216696, 408011092944, 533658335895, 715687728967]\n",
    "    else:\n",
    "        raise Exception(\"unknown CONFIG\")\n",
    "else:\n",
    "    raise Exception(\"unknown SF\")\n",
    "\n",
    "BYTE_FACTOR = 1e6\n",
    "SECOND_FACTOR = 1e9\n",
    "    \n",
    "bytes_used = [int(x / BYTE_FACTOR) for x in bytes_used]\n",
    "times = [int(x / SECOND_FACTOR) for x in times]\n",
    "phase_ends = [int(x / SECOND_FACTOR) for x in phase_ends]\n",
    "\n",
    "bytes_used_delete = [int(x / BYTE_FACTOR) for x in bytes_used_delete]\n",
    "times_delete = [int(x / SECOND_FACTOR) for x in times_delete]\n",
    "phase_ends_delete = [int(x / SECOND_FACTOR) for x in phase_ends_delete]\n",
    "\n",
    "bytes_used_delete_encoded = [int(x / BYTE_FACTOR) for x in bytes_used_delete_encoded]\n",
    "times_delete_encoded = [int(x / SECOND_FACTOR) for x in times_delete_encoded]\n",
    "phase_ends_delete_encoded = [int(x / SECOND_FACTOR) for x in phase_ends_delete_encoded]\n",
    "\n",
    "lineitem_dictionary_bytes = int(lineitem_dictionary_bytes / BYTE_FACTOR)\n",
    "lineitem_value_bytes = int(lineitem_value_bytes / BYTE_FACTOR)\n",
    "\n",
    "labels = ['Without physical delete', 'With physical delete', 'With physical delete and encoding']\n",
    "x_values = [times, times_delete, times_delete_encoded]\n",
    "y_values = [bytes_used, bytes_used_delete, bytes_used_delete_encoded]\n",
    "vline_values = [phase_ends, phase_ends_delete, phase_ends_delete_encoded]\n",
    "\n",
    "#fig, axs = plt.subplots(3, figsize=(5,10), sharex=True, sharey=True)\n",
    "fig, ax = plt.subplots()\n",
    "    \n",
    "ax.set_xlabel('Clustering run time (seconds)')\n",
    "ax.set_ylabel('Memory consumption of lineitem (MB)')\n",
    "ax.get_yaxis().set_major_formatter(matplotlib.ticker.FuncFormatter(lambda x, p: format(int(x), ',')))\n",
    "\n",
    "for v in range(3):\n",
    "    ax.plot(x_values[v], y_values[v], label=labels[v], linestyle='dotted', marker='.')\n",
    "    max_memory = max(y_values[v])\n",
    "\n",
    "    steps = ['partition', 'merge', 'sort', 'encode']\n",
    "    linestyles = ['dotted', '', 'dashed', 'dashdot']\n",
    "    for index, phase_end in enumerate(vline_values[v]):\n",
    "        if index == 1:\n",
    "            # do not plot merge line\n",
    "            continue\n",
    "        time_index = bisect.bisect(x_values[v], phase_end)\n",
    "        byte_low = y_values[v][time_index] - 3000\n",
    "        byte_high = y_values[v][time_index] + 3000\n",
    "\n",
    "        \n",
    "        if v == 0:\n",
    "            ax.vlines(phase_end, byte_low, byte_high, label=f'End {steps[index]} step', linestyle=linestyles[index], linewidth=2 if index == 0 else 1)\n",
    "        else:\n",
    "            ax.vlines(phase_end, byte_low, byte_high, linestyle=linestyles[index], linewidth=2 if index == 0 else 1)\n",
    "\n",
    "\n",
    "max_x_values = max(max(x_values[0]), max(x_values[1]), max(x_values[2]) )\n",
    "ax.hlines(lineitem_dictionary_bytes, 0, max_x_values, label='lineitem with dictionary segments', linestyle='dashed', color='gray', linewidth=1)\n",
    "ax.hlines(lineitem_value_bytes, 0, max_x_values, label='lineitem with value segments', linestyle='dotted', color='gray', linewidth=2)\n",
    "\n",
    "PAPER_MODE = True\n",
    "if PAPER_MODE:\n",
    "    box = ax.get_position()\n",
    "    #ax.set_position([box.x0, box.y0, box.width, box.height * 0.8])\n",
    "    legend = ax.legend(loc=\"lower center\", ncol=2, bbox_to_anchor=(0.5, 1))\n",
    "else:\n",
    "    legend = ax.legend(loc='center left', ncol=1, bbox_to_anchor=(1.0, 0.5))\n",
    "\n",
    "plt.savefig(f'/home/aloeser/Downloads/memory_consumption{CONFIG}.pdf', bbox_inches='tight', bbox_extra_artists=(legend,))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}