4. Index对象的创建，查、改、增、删和使用.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Index对象的创建,、查、改、增、删和使用\n",
    "想要用好pandas，必须了解其核心对象之一的**索引**。\n",
    "- 索引类似于元组，其本身是不能赋值修改的；\n",
    "- 其在数据进行整体运算时，辅助自动对齐，这是pandas不同于其他数据处理库的一大特征；\n",
    "- 多层索引可以帮助改变表的形态，如透视表等。\n",
    "\n",
    "所以，这一章要仔细学习。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "__auther__ = 'zhenhang.sun@gmail.com'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:41.317508Z",
     "start_time": "2017-09-25T14:27:41.297966Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'D:\\\\github\\\\pandas-tutorial'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:41.927092Z",
     "start_time": "2017-09-25T14:27:41.320505Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. 单层索引"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.1 创建\n",
    "##### `pd.Index(data, dtype=Object, name=None)`\n",
    "- name：一维列表\n",
    "- dtype：索引元素的类型，默认为object型\n",
    "- name：索引的名字，类似于列的名字"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:41.951528Z",
     "start_time": "2017-09-25T14:27:41.931514Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c'], dtype='object', name='name1')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = ['a','b','c']\n",
    "index = pd.Index(data, name = 'name1')\n",
    "index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**从返回值可以看到，index由三部分组成，可以分别查看。**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.076234Z",
     "start_time": "2017-09-25T14:27:41.955533Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'name1'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.219942Z",
     "start_time": "2017-09-25T14:27:42.081237Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['a', 'b', 'c'], dtype=object)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.355059Z",
     "start_time": "2017-09-25T14:27:42.223945Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('O')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.dtype"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.2 查\n",
    "- 查询方式和一维ndarray或Series的.iloc[]完全一样。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.501178Z",
     "start_time": "2017-09-25T14:27:42.358062Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'a'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index[0] # scalar，返回值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.567237Z",
     "start_time": "2017-09-25T14:27:42.503180Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b'], dtype='object', name='name1')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index[0:2] # 范围，返回index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.687335Z",
     "start_time": "2017-09-25T14:27:42.570239Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'c'], dtype='object', name='name1')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index[[0, 2]] # 列表，返回index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.786413Z",
     "start_time": "2017-09-25T14:27:42.690338Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'c'], dtype='object', name='name1')"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = [True, False, True]  # mask,返回index\n",
    "index[mask]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.3 改索引名\n",
    "虽然索引的值是不能修改的，但是名字确是可以修改的。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.3.1 直接改"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.890496Z",
     "start_time": "2017-09-25T14:27:42.789415Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.name = 'new_name' \n",
    "index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.3.2 函数改\n",
    "##### `Index.set_names(names, inplace=False)`\n",
    "- names：要设置的名字，可以为名字的列表；\n",
    "- inplace：是否原地修改。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:42.989576Z",
     "start_time": "2017-09-25T14:27:42.896498Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.set_names('new_name')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.4 增"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.4.1 按位置添加一行\n",
    "##### `Index.insert(loc, value)`\n",
    "- loc：位置编号\n",
    "- value：值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.089654Z",
     "start_time": "2017-09-25T14:27:42.991578Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.186730Z",
     "start_time": "2017-09-25T14:27:43.091655Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'd', 'b', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.insert(1,'d')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.4.2 尾部添加多行\n",
    "##### `Index.append(other)`\n",
    "- other：其他索引对象"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.274797Z",
     "start_time": "2017-09-25T14:27:43.189733Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index1 = index.copy()\n",
    "index1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.377884Z",
     "start_time": "2017-09-25T14:27:43.281815Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c', 'a', 'b', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index1.append(index)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.4.2 并\n",
    "##### `Index.union(other)`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.473890Z",
     "start_time": "2017-09-25T14:27:43.382880Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['b', 'c', 'd'], dtype='object')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index2 = pd.Index(['b','c','d'])\n",
    "index2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.602988Z",
     "start_time": "2017-09-25T14:27:43.475893Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'b', 'c', 'd'], dtype='object')"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index1.union(index2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.5 删"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.5.1 按位置删除一行\n",
    "##### `Index.delete(loc)`\n",
    "- loc：位置编号"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.697057Z",
     "start_time": "2017-09-25T14:27:43.605990Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index1.delete(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.5.2 按索引删除多行\n",
    "##### `Index.drop(labels)`\n",
    "- labels：索引列表"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.789129Z",
     "start_time": "2017-09-25T14:27:43.700060Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['c'], dtype='object', name='new_name')"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index1.drop(['a','b'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.5.3 取交\n",
    "##### `Index.intersection(other)`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:43.887204Z",
     "start_time": "2017-09-25T14:27:43.792131Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['b', 'c'], dtype='object')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index1.intersection(index2) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. 多层索引"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.1 创建\n",
    "##### `pd.MultiIndex.from_tuples(labels, names=None)`\n",
    "- labels：元组或列表的列表；\n",
    "- names：名字的列表。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.033933Z",
     "start_time": "2017-09-25T14:27:43.890207Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['a', 'b'], ['one', 'two']],\n",
       "           labels=[[0, 0, 1], [0, 1, 0]],\n",
       "           names=['name1', 'name2'])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# data = [['a','one'],['a','two'],['b','one']]\n",
    "data = [('a','one'),('a','two'),('b','one')]\n",
    "index = pd.MultiIndex.from_tuples(data, names=['name1','name2'])\n",
    "index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.156527Z",
     "start_time": "2017-09-25T14:27:44.036936Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name1  name2\n",
       "a      one      1\n",
       "       two      2\n",
       "b      one      3\n",
       "dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s = pd.Series([1,2,3], index = index)\n",
    "s"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.2 查\n",
    "- 查询方法和单层索引完全一致。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.256602Z",
     "start_time": "2017-09-25T14:27:44.161548Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('a', 'one')"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index[0]  # scalar，返回值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.357680Z",
     "start_time": "2017-09-25T14:27:44.258605Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['a', 'b'], ['one', 'two']],\n",
       "           labels=[[0, 0], [0, 1]],\n",
       "           names=['name1', 'name2'])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index[0:2]  # 范围，返回MultiIndex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.456400Z",
     "start_time": "2017-09-25T14:27:44.363684Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['a', 'b'], ['one', 'two']],\n",
       "           labels=[[0, 1], [0, 0]],\n",
       "           names=['name1', 'name2'])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index[[0,2]] # 列表，返回MultiIndex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.584996Z",
     "start_time": "2017-09-25T14:27:44.461417Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['a', 'b'], ['one', 'two']],\n",
       "           labels=[[0, 1], [0, 0]],\n",
       "           names=['name1', 'name2'])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = [True, False, True] # mask，返回MultiIndex\n",
    "index[mask]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 获取某一层索引 MultiIndex.get_level_values(level)\n",
    "- level：int，选中的那一层"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.690075Z",
     "start_time": "2017-09-25T14:27:44.587999Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['a', 'a', 'b'], dtype='object', name='name1')"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.get_level_values(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.787150Z",
     "start_time": "2017-09-25T14:27:44.694089Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['one', 'two', 'one'], dtype='object', name='name2')"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.get_level_values(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.3 改"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3.1 改索引名\n",
    "##### `MultiIndex.set_names(names, level=None, inplace=False)`\n",
    "- names：要设置的名字，可以为名字的列表；\n",
    "- level：多层索引需要设置修改的索引层次，可以为列表，要与names匹配；\n",
    "- inplace：是否原地修改。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:44.953874Z",
     "start_time": "2017-09-25T14:27:44.789151Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['a', 'b'], ['one', 'two']],\n",
       "           labels=[[0, 0, 1], [0, 1, 0]],\n",
       "           names=['new_name_1', 'name2'])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.set_names('new_name_1',level=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3.2 改索引层次顺序\n",
    "##### `MultiIndex.swaplevel(i=-2, j=-1)`\n",
    "- 改变level i 和level j的次序"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:45.102482Z",
     "start_time": "2017-09-25T14:27:44.955877Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['one', 'two'], ['a', 'b']],\n",
       "           labels=[[0, 1, 0], [0, 0, 1]],\n",
       "           names=['name2', 'name1'])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index.swaplevel()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### `Series.swaplevel(i=-2, j=-1)`\n",
    "#####  `DataFrame.swaplevel(i=-2, j=-1, axis=1)`\n",
    "- axis：0-行索引，1-列索引。\n",
    "这两个函数更实用一些。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:45.276627Z",
     "start_time": "2017-09-25T14:27:45.105487Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name2  name1\n",
       "one    a        1\n",
       "two    a        2\n",
       "one    b        3\n",
       "dtype: int64"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s.swaplevel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:45.424732Z",
     "start_time": "2017-09-25T14:27:45.280620Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns = index.copy()\n",
    "columns.set_names( names = ['name3','name4'], level = [0,1], inplace = True) #列索引取和行索引相同，只是改了名字\n",
    "df = pd.DataFrame([[1,2,3],[4,5,6],[7,8,9]], index= index, columns = columns)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:45.554384Z",
     "start_time": "2017-09-25T14:27:45.429736Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th>a</th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name4       one two one\n",
       "name3         a   a   b\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.swaplevel(axis=1)  # 交换列索引顺序"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. 多层索引使用方法\n",
    "当对values进行查看时，多层索引可以分开使用。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:45.671972Z",
     "start_time": "2017-09-25T14:27:45.559393Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = df.copy()\n",
    "df1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**索引为空不代表缺失，缺省写法，意思是之前的索引一致。**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.1 对于外层索引\n",
    "**记住：**\n",
    "- 无论是Series还是DataFrame，外层索引都是可以直接使用，也就是说可以认为只有这一层索引；\n",
    "- **用法和第二篇查、改、增、删提到的方法完全相同**。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1.1 []\n",
    "快捷操作，还是四种用法。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:45.931171Z",
     "start_time": "2017-09-25T14:27:45.674985Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name4        one\n",
       "name1 name2     \n",
       "a     one      3\n",
       "      two      6\n",
       "b     one      9"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1['b']  # 列外层"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.033253Z",
     "start_time": "2017-09-25T14:27:45.935179Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1[['a','b']] # 列外层"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.139341Z",
     "start_time": "2017-09-25T14:27:46.036263Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1[0:2]  # 行外层"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.244377Z",
     "start_time": "2017-09-25T14:27:46.144345Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask =[True, False, True]  # 行外层\n",
    "df1[mask]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1.2  .loc[]\n",
    "行和列索引都使用索引形式。\n",
    "\n",
    "**下面都以第一维度为例，第二维可以类比。**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.429017Z",
     "start_time": "2017-09-25T14:27:46.248384Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>one</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name4  one\n",
       "name2     \n",
       "one      3\n",
       "two      6"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.loc['a','b']  # 单行索引'a' "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.532097Z",
     "start_time": "2017-09-25T14:27:46.431018Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name4        one\n",
       "name1 name2     \n",
       "a     one      3\n",
       "      two      6\n",
       "b     one      9"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.loc['a':'b', 'b']   #范围'a': 'b'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.661701Z",
     "start_time": "2017-09-25T14:27:46.534599Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name4        one\n",
       "name1 name2     \n",
       "a     one      3\n",
       "      two      6\n",
       "b     one      9"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.loc[['a','b'], 'b'] #列表"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.753770Z",
     "start_time": "2017-09-25T14:27:46.664703Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <th>one</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name4        one\n",
       "name1 name2     \n",
       "a     one      3\n",
       "b     one      9"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = [True, False, True] # mask\n",
    "df1.loc[mask,'b']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1.3 .iloc[]\n",
    "这个简单，可以把索引都忽略掉，行和列都使用位置形式。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.885867Z",
     "start_time": "2017-09-25T14:27:46.758784Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name3  name4\n",
       "a      one      1\n",
       "       two      2\n",
       "Name: (a, one), dtype: int64"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.iloc[0,0:2] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:46.995950Z",
     "start_time": "2017-09-25T14:27:46.887869Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a    \n",
       "name4       one two\n",
       "name1 name2        \n",
       "a     one     1   2\n",
       "      two     4   5"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.iloc[0:2, 0:2] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.094026Z",
     "start_time": "2017-09-25T14:27:46.998952Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a    \n",
       "name4       one two\n",
       "name1 name2        \n",
       "a     one     1   2\n",
       "      two     4   5"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.iloc[[0,1],0:2] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.199106Z",
     "start_time": "2017-09-25T14:27:47.096027Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a    \n",
       "name4       one two\n",
       "name1 name2        \n",
       "a     one     1   2\n",
       "b     one     7   8"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = [True, False, True]\n",
    "df1.iloc[mask,0:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2 对于内层索引\n",
    "- **内层索引不可直接使用，必须先外层、再内层，直接使用会报错；**\n",
    "- 内层只能使用单索引形式，其他形式报错。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2.1 [ , ]\n",
    "快捷操作，只有一种用法，取出一列。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.305853Z",
     "start_time": "2017-09-25T14:27:47.203108Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.400425Z",
     "start_time": "2017-09-25T14:27:47.312857Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name1  name2\n",
       "a      one      1\n",
       "       two      4\n",
       "b      one      7\n",
       "Name: (a, one), dtype: int64"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1['a','one']  #取一列，先外层单列索引，再内层单列索引，其他形式都报错"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### 3.2.2 .loc[ , ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.507512Z",
     "start_time": "2017-09-25T14:27:47.402428Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "name3  name4\n",
       "a      one      1\n",
       "       two      2\n",
       "b      one      3\n",
       "Name: (a, one), dtype: int64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.loc['a','one'] # 取一行，先外层单行索引，再内层单列索引，其他形式都报错"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2.3 .iloc[ , ]\n",
    "这种方法不受影响，因为 .iloc[] 无视索引，只按照位置定位，所以和3.1.3节外层索引部分完全相同。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.3 xs直接选取法\n",
    "适合在单层level选取，不能行列同时操作。\n",
    "##### `Series.xs(key, level=None, drop_level=True)`\n",
    "##### `DataFrame.xs(key, axis=0, level=None, drop_level=True)`\n",
    "- key: 要选取的索引值或其列表；\n",
    "- axis：0-行索引，1-列索引；\n",
    "- level：索引层次；\n",
    "- drop_level：True or False，是否显示用于选取的level索引，默认不显示。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.638613Z",
     "start_time": "2017-09-25T14:27:47.513511Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3         a       b\n",
       "name4       one two one\n",
       "name1 name2            \n",
       "a     one     1   2   3\n",
       "      two     4   5   6\n",
       "b     one     7   8   9"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = df.copy()\n",
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.762705Z",
     "start_time": "2017-09-25T14:27:47.646619Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>name3</th>\n",
       "      <th colspan=\"2\" halign=\"left\">a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name4</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>one</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3   a       b\n",
       "name4 one two one\n",
       "name1            \n",
       "a       1   2   3\n",
       "b       7   8   9"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.xs('one', axis=0, level=1)  # 行索引的level 1, 有两行"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-25T14:27:47.859778Z",
     "start_time": "2017-09-25T14:27:47.765707Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name3</th>\n",
       "      <th>a</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name1</th>\n",
       "      <th>name2</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>one</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>one</th>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "name3        a\n",
       "name1 name2   \n",
       "a     one    2\n",
       "      two    5\n",
       "b     one    8"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.xs('two', axis=1, level=1) # 列索引的level 1，有一列"
   ]
  },
  {
   "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": "3.7.0"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {
    "height": "300px",
    "width": "252px"
   },
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "485px",
    "left": "0px",
    "right": "1089px",
    "top": "66px",
    "width": "191px"
   },
   "toc_section_display": "block",
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}