notebook/01-Pandas数据结构详解.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> 教你学会 Pandas 不是我的目的，**教你轻松玩转 Pandas 才是我的目的**。我会通过一系列实例来带入 Pandas 的知识点，让你在学习 Pandas 的路上不再枯燥。\n",
    "\n",
    "> 声明：我所写的**轻松玩转 Pandas 教程都是免费的**，如果对你有帮助，你可以持续关注我。\n",
    "\n",
    "Pandas 有很多高级的功能，但是想要掌握高级功能前，需要先掌握它的基础知识，Pandas 中的数据结构算是非常基础的知识之一了。\n",
    "\n",
    "Pandas 常用的数据结构有两种：Series 和 DataFrame。这些数据结构构建在 Numpy 数组之上，这意味着它们效率很高。我们来分别看看这些数据结构都长什么样子吧。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:01.781415Z",
     "start_time": "2018-07-09T13:01:01.776401Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 导入相关库\n",
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Series\n",
    "\n",
    "### 简介\n",
    "\n",
    "Series 是一个带有 **名称** 和索引的一维数组，既然是数组，肯定要说到的就是数组中的元素类型，在 Series 中包含的数据类型可以是整数、浮点、字符串、Python对象等。\n",
    "\n",
    "假定有一个场景是：存储一些用户的信息，暂时只包括年龄信息。\n",
    "\n",
    "我们可以通过 Series 来存储，这里我们通过 Series 存储了四个年龄：18/30/25/40，**只需将要存储的数据构建成一个数组，然后赋值给data参数即可**。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:01.842580Z",
     "start_time": "2018-07-09T13:01:01.785426Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    18\n",
       "1    30\n",
       "2    25\n",
       "3    40\n",
       "dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 存储了 4 个年龄：18/30/25/40\n",
    "user_age = pd.Series(data=[18, 30, 25, 40])\n",
    "user_age"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-04T02:52:44.335318Z",
     "start_time": "2018-06-04T02:52:44.289195Z"
    }
   },
   "source": [
    "可以看到，已经正确将多个年龄存储到 Series 中了，你可能会想，单独存储了年龄有什么用，我怎么知道这个年龄属于哪个用户呢？\n",
    "\n",
    "我们可以通过 Series 的 index（索引）来解决这个问题。由于有四个年龄，自然地也需要四个姓名，所以我们**需要构建一个与 data 长度相同的数组**，然后通过下面的操作即可满足要求。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:01.875666Z",
     "start_time": "2018-07-09T13:01:01.846589Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tom      18\n",
       "Bob      30\n",
       "Mary     25\n",
       "James    40\n",
       "dtype: int64"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_age.index = [\"Tom\", \"Bob\", \"Mary\", \"James\"]\n",
    "user_age"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "你看，现在姓名与年龄已经完全对应上了。虽然说我们自己知道 Tom/Bob 这些是姓名，但是别人不知道啊，我们怎么告诉他人呢？\n",
    "\n",
    "要想让别人知道，我们可以为 index 起个名字。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:01.910759Z",
     "start_time": "2018-07-09T13:01:01.880680Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      18\n",
       "Bob      30\n",
       "Mary     25\n",
       "James    40\n",
       "dtype: int64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_age.index.name = \"name\"\n",
    "user_age"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-04T03:02:31.869982Z",
     "start_time": "2018-06-04T03:02:31.861996Z"
    }
   },
   "source": [
    "可能你还会想，如果别人在看我写的代码，怎么能快速的知道我这写的到底是什么玩意呢？\n",
    "\n",
    "别急，就像我们给index起名字一样，我们也可以给 Series 起个名字。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:01.985960Z",
     "start_time": "2018-07-09T13:01:01.917779Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      18\n",
       "Bob      30\n",
       "Mary     25\n",
       "James    40\n",
       "Name: user_age_info, dtype: int64"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_age.name=\"user_age_info\"\n",
    "user_age"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "通过上面一系列的操作，我们对 Series 的结构上有了基本的了解，简单来说，**一个 Series 包括了 data、index 以及 name**。\n",
    "\n",
    "上面的操作非常方便做演示来使用，如果想要快速实现上面的功能，可以通过以下方式来实现。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.042110Z",
     "start_time": "2018-07-09T13:01:01.991976Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      18\n",
       "Bob      30\n",
       "Mary     25\n",
       "James    40\n",
       "Name: user_age_info, dtype: int64"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 构建索引\n",
    "name = pd.Index([\"Tom\", \"Bob\", \"Mary\", \"James\"], name=\"name\")\n",
    "# 构建 Series\n",
    "user_age = pd.Series(data=[18, 30, 25, 40], index=name, name=\"user_age_info\")\n",
    "user_age"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "另外，需要说明的是我们在构造 Series 的时候，并没有设定每个元素的数据类型，这个时候，pandas 会自动判断一个数据类型，并作为 Series 的类型。\n",
    "\n",
    "当然了，我们也可以自己手动指定数据类型。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.080211Z",
     "start_time": "2018-07-09T13:01:02.047124Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      18.0\n",
       "Bob      30.0\n",
       "Mary     25.0\n",
       "James    40.0\n",
       "Name: user_age_info, dtype: float64"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 指定类型为浮点型\n",
    "user_age = pd.Series(data=[18, 30, 25, 40], index=name, name=\"user_age_info\", dtype=float)\n",
    "user_age"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Series 像什么\n",
    "\n",
    "Series 包含了 dict 的特点，也就意味着可以使用与 dict 类似的一些操作。我们可以将 index 中的元素看成是 dict 中的 key。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.117310Z",
     "start_time": "2018-07-09T13:01:02.088232Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18.0"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 获取 Tom 的年龄\n",
    "user_age[\"Tom\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-04T03:42:39.368547Z",
     "start_time": "2018-06-04T03:42:39.358520Z"
    }
   },
   "source": [
    "此外，可以通过 get 方法来获取。通过这种方式的好处是当索引不存在时，不会抛出异常。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.148393Z",
     "start_time": "2018-07-09T13:01:02.123326Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18.0"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_age.get(\"Tom\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Series 除了像 dict 外，也非常像 ndarray，这也就意味着可以采用切片操作。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.187497Z",
     "start_time": "2018-07-09T13:01:02.155412Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18.0"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 获取第一个元素\n",
    "user_age[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.244651Z",
     "start_time": "2018-07-09T13:01:02.192511Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom     18.0\n",
       "Bob     30.0\n",
       "Mary    25.0\n",
       "Name: user_age_info, dtype: float64"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 获取前三个元素\n",
    "user_age[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.296789Z",
     "start_time": "2018-07-09T13:01:02.248661Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "James    40.0\n",
       "Name: user_age_info, dtype: float64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 获取年龄大于30的元素\n",
    "user_age[user_age > 30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.341909Z",
     "start_time": "2018-07-09T13:01:02.303808Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "James    40.0\n",
       "Bob      30.0\n",
       "Name: user_age_info, dtype: float64"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 获取第4个和第二个元素\n",
    "user_age[[3, 1]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "可以看到，无论我们通过切片如何操作 Series ，它都能够自动对齐 index。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Series 的向量化操作\n",
    "\n",
    "Series 与 ndarray 一样，也是支持向量化操作的。同时也可以传递给大多数期望 ndarray 的 NumPy 方法。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.388032Z",
     "start_time": "2018-07-09T13:01:02.344921Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      19.0\n",
       "Bob      31.0\n",
       "Mary     26.0\n",
       "James    41.0\n",
       "Name: user_age_info, dtype: float64"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_age + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.450198Z",
     "start_time": "2018-07-09T13:01:02.392044Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      6.565997e+07\n",
       "Bob      1.068647e+13\n",
       "Mary     7.200490e+10\n",
       "James    2.353853e+17\n",
       "Name: user_age_info, dtype: float64"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(user_age)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DataFrame\n",
    "\n",
    "DataFrame 是一个带有**索引**的二维数据结构，每列可以有自己的名字，并且可以有不同的数据类型。你可以把它想象成一个 excel 表格或者数据库中的一张表，DataFrame 是最常用的 pandas 对象。\n",
    "\n",
    "我们继续使用之前的实例来讲解 DataFrame，在存储用户信息时，除了年龄之外，我还想存储用户所在的城市。如何通过 DataFrame 实现呢？\n",
    "\n",
    "可以构建一个 dict，key 是需要存储的信息，value 是信息列表。**然后将 dict 传递给 data 参数**。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.511361Z",
     "start_time": "2018-07-09T13:01:02.457217Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city\n",
       "name                 \n",
       "Tom     18    BeiJing\n",
       "Bob     30   ShangHai\n",
       "Mary    25  GuangZhou\n",
       "James   40   ShenZhen"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index = pd.Index(data=[\"Tom\", \"Bob\", \"Mary\", \"James\"], name=\"name\")\n",
    "\n",
    "data = {\n",
    "    \"age\": [18, 30, 25, 40],\n",
    "    \"city\": [\"BeiJing\", \"ShangHai\", \"GuangZhou\", \"ShenZhen\"]\n",
    "}\n",
    "\n",
    "user_info = pd.DataFrame(data=data, index=index)\n",
    "user_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "可以看到，我们成功构建了一个 DataFrame，这个 DataFrame 的索引是用户性别，还有两列分别是用户的年龄和城市信息。\n",
    "\n",
    "除了上面这种传入 dict 的方式构建外，我们还可以通过另外一种方式来构建。这种方式是**先构建一个二维数组，然后再生成一个列名称列表**。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.558487Z",
     "start_time": "2018-07-09T13:01:02.515373Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city\n",
       "name                 \n",
       "Tom     18    BeiJing\n",
       "Bob     30   ShangHai\n",
       "Mary    25  GuangZhou\n",
       "James   40   ShenZhen"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = [[18, \"BeiJing\"], \n",
    "        [30, \"ShangHai\"], \n",
    "        [25, \"GuangZhou\"], \n",
    "        [40, \"ShenZhen\"]]\n",
    "columns = [\"age\", \"city\"]\n",
    "\n",
    "user_info = pd.DataFrame(data=data, index=index, columns=columns)\n",
    "user_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 访问行\n",
    "\n",
    "在生成了 DataFrame 之后，可以看到，每一行就表示某一个用户的信息，假如我想要访问 Tom 的信息，我该如何操作呢？\n",
    "\n",
    "一种办法是通过索引名来访问某行，这种办法需要借助 loc 方法。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.608620Z",
     "start_time": "2018-07-09T13:01:02.562498Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age          18\n",
       "city    BeiJing\n",
       "Name: Tom, dtype: object"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.loc[\"Tom\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "除了直接通过索引名来访问某一行数据之外，还可以通过这行所在的位置来选择这一行。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.668781Z",
     "start_time": "2018-07-09T13:01:02.611629Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age          18\n",
       "city    BeiJing\n",
       "Name: Tom, dtype: object"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.iloc[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-04T07:04:11.777185Z",
     "start_time": "2018-06-04T07:04:11.751116Z"
    }
   },
   "source": [
    "现在能够访问某一个用户的信息了，那么我如何访问多个用户的信息呢？也就是如何访问多行呢？\n",
    "\n",
    "借助行切片可以轻松完成，来看这里。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.703875Z",
     "start_time": "2018-07-09T13:01:02.671790Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      age       city\n",
       "name                \n",
       "Bob    30   ShangHai\n",
       "Mary   25  GuangZhou"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.iloc[1:3]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 访问列\n",
    "\n",
    "学会了如何访问行数据之外，自然而然会想到如何访问列。我们可以通过属性（“.列名”）的方式来访问该列的数据，也可以通过[column]的形式来访问该列的数据。\n",
    "\n",
    "假如我想获取所有用户的年龄，那么可以这样操作。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.756014Z",
     "start_time": "2018-07-09T13:01:02.710893Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      18\n",
       "Bob      30\n",
       "Mary     25\n",
       "James    40\n",
       "Name: age, dtype: int64"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.age"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.808153Z",
     "start_time": "2018-07-09T13:01:02.761027Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name\n",
       "Tom      18\n",
       "Bob      30\n",
       "Mary     25\n",
       "James    40\n",
       "Name: age, dtype: int64"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info[\"age\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "如果想要同时获取年龄和城市该如何操作呢？"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.859289Z",
     "start_time": "2018-07-09T13:01:02.816174Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>city</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>BeiJing</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>ShangHai</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>GuangZhou</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>ShenZhen</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            city  age\n",
       "name                 \n",
       "Tom      BeiJing   18\n",
       "Bob     ShangHai   30\n",
       "Mary   GuangZhou   25\n",
       "James   ShenZhen   40"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 可以变换列的顺序\n",
    "user_info[[\"city\", \"age\"]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-04T06:10:27.665304Z",
     "start_time": "2018-06-04T06:10:27.641243Z"
    }
   },
   "source": [
    "### 新增/删除列\n",
    "\n",
    "在生成了 DataFrame 之后，突然你发现好像缺失了用户的性别这个信息，那么如何添加呢？\n",
    "\n",
    "如果所有的性别都一样，我们可以通过传入一个标量，pandas 会自动帮我们广播来填充所有的位置。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.899396Z",
     "start_time": "2018-07-09T13:01:02.864303Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "      <th>sex</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city   sex\n",
       "name                       \n",
       "Tom     18    BeiJing  male\n",
       "Bob     30   ShangHai  male\n",
       "Mary    25  GuangZhou  male\n",
       "James   40   ShenZhen  male"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info[\"sex\"] = \"male\"\n",
    "user_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "如果想要删除某一列，可以使用 pop 方法来完成。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:02.969583Z",
     "start_time": "2018-07-09T13:01:02.903407Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city\n",
       "name                 \n",
       "Tom     18    BeiJing\n",
       "Bob     30   ShangHai\n",
       "Mary    25  GuangZhou\n",
       "James   40   ShenZhen"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.pop(\"sex\")\n",
    "user_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-04T06:36:25.481870Z",
     "start_time": "2018-06-04T06:36:25.473849Z"
    }
   },
   "source": [
    "如果用户的性别不一致的时候，我们可以通过传入一个 like-list 来添加新的一列。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:03.030747Z",
     "start_time": "2018-07-09T13:01:02.978607Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "      <th>sex</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "      <td>female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "      <td>male</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city     sex\n",
       "name                         \n",
       "Tom     18    BeiJing    male\n",
       "Bob     30   ShangHai    male\n",
       "Mary    25  GuangZhou  female\n",
       "James   40   ShenZhen    male"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info[\"sex\"] = [\"male\", \"male\", \"female\", \"male\"]\n",
    "user_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "通过上面的例子可以看出，我们创建新的列的时候都是在原有的 DataFrame 上修改的，也就是说如果添加了新的一列之后，原有的 DataFrame 会发生改变。\n",
    "\n",
    "如果想要保证原有的 DataFrame 不改变的话，我们可以通过 assign 方法来创建新的一列。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:03.082885Z",
     "start_time": "2018-07-09T13:01:03.036762Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "      <th>sex</th>\n",
       "      <th>age_add_one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "      <td>male</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "      <td>male</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "      <td>female</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "      <td>male</td>\n",
       "      <td>41</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city     sex  age_add_one\n",
       "name                                      \n",
       "Tom     18    BeiJing    male           19\n",
       "Bob     30   ShangHai    male           31\n",
       "Mary    25  GuangZhou  female           26\n",
       "James   40   ShenZhen    male           41"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.assign(age_add_one = user_info[\"age\"] + 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T13:01:03.128007Z",
     "start_time": "2018-07-09T13:01:03.090908Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>city</th>\n",
       "      <th>sex</th>\n",
       "      <th>sex_code</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Tom</th>\n",
       "      <td>18</td>\n",
       "      <td>BeiJing</td>\n",
       "      <td>male</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bob</th>\n",
       "      <td>30</td>\n",
       "      <td>ShangHai</td>\n",
       "      <td>male</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mary</th>\n",
       "      <td>25</td>\n",
       "      <td>GuangZhou</td>\n",
       "      <td>female</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>James</th>\n",
       "      <td>40</td>\n",
       "      <td>ShenZhen</td>\n",
       "      <td>male</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age       city     sex  sex_code\n",
       "name                                   \n",
       "Tom     18    BeiJing    male         1\n",
       "Bob     30   ShangHai    male         1\n",
       "Mary    25  GuangZhou  female         0\n",
       "James   40   ShenZhen    male         1"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_info.assign(sex_code = np.where(user_info[\"sex\"] == \"male\", 1, 0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------\n",
    "\n",
    "\n",
    "想要学习更多关于人工智能的知识，请关注公众号：**AI派**\n",
    "\n",
    "![qrcode_for_gh_60cef389e81c_258.jpg](http://www.naodongopen.com/wp-content/uploads/2018/06/qrcode_for_gh_60cef389e81c_258.jpg)\n",
    "这里我将整篇文章的内容整理成了pdf，想要pdf文件的可以在公众号后台回复关键字：**pandas01**。"
   ]
  }
 ],
 "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}