{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CHAPTER 11 Time Series（时间序列）\n",
    "\n",
    "时间序列指能在任何能在时间上观测到的数据。很多时间序列是有固定频率（fixed frequency）的，意思是数据点会遵照某种规律定期出现，比如每15秒，每5分钟，或每个月。时间序列也可能是不规律的（irregular），没有一个固定的时间规律。如何参照时间序列数据取决于我们要做什么样的应用，我们可能会遇到下面这些：\n",
    "\n",
    "- Timestamps（时间戳），具体的某一个时刻\n",
    "- Fixed periods（固定的时期），比如2007年的一月，或者2010年整整一年\n",
    "- Intervals of time（时间间隔），通常有一个开始和结束的时间戳。Periods（时期）可能被看做是Intervals（间隔）的一种特殊形式。\n",
    "- Experiment or elapsed time（实验或经过的时间）；每一个时间戳都是看做是一个特定的开始时间（例如，在放入烤箱后，曲奇饼的直径在每一秒的变化程度）\n",
    "\n",
    "这一章主要涉及前三个类型。\n",
    "\n",
    "> pandas也支持基于timedeltas的index，本书不会对timedelta index做介绍，感兴趣的可以查看pandas的文档。\n",
    "\n",
    "\n",
    "# 11.1 Date and Time Data Types and Tools（日期和时间数据类型及其工具）\n",
    "\n",
    "python有标准包用来表示时间和日期数据。datetime, time, calendar，这些模块经常被使用。datetime.datetime类型，或简单写为datetime，被广泛使用：\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from datetime import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "now = datetime.now()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2017, 12, 1, 12, 12, 0, 375896)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "now"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2017, 12, 1)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "now.year, now.month, now.day"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "datetime能保存日期和时间到微妙级别。timedelta表示两个不同的datetime对象之间的时间上的不同："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.timedelta(926, 56700)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "delta = datetime(2011, 1, 7) - datetime(2008, 6, 24, 8, 15)\n",
    "delta"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "926"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "delta.days"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "56700"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "delta.seconds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "我们可以在一个datetime对象上，添加或减少一个或多个timedelta，这样可以产生新的变化后的对象："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from datetime import timedelta"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "start = datetime(2011, 1, 7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2011, 1, 19, 0, 0)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "start + timedelta(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2010, 12, 14, 0, 0)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "start - 2 * timedelta(12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "下表汇总了一些datetime模块中的数据类型：\n",
    "\n",
    "![](http://oydgk2hgw.bkt.clouddn.com/pydata-book/wqo6m.png)\n",
    "\n",
    "# 1 Converting Between String and Datetime（字符串与时间的转换）\n",
    "\n",
    "我们可以对datetime对象，以及pandas的Timestamp对象进行格式化，这部分之后会介绍，使用str或strftime方法，传入一个特定的时间格式就能进行转换："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "stamp = datetime(2011, 1, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2011-01-03 00:00:00'"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "str(stamp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2011-01-03'"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stamp.strftime('%Y-%m-%d')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "下表是关于日期时间类型的格式：\n",
    "\n",
    "![](http://oydgk2hgw.bkt.clouddn.com/pydata-book/r98dw.png)\n",
    "\n",
    "![](http://oydgk2hgw.bkt.clouddn.com/pydata-book/bc9e8.png)\n",
    "\n",
    "我们可以利用上面的format codes（格式码；时间日期格式）把字符串转换为日期，这要用到datetime.strptime:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "value = '2011-01-03'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2011, 1, 3, 0, 0)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "datetime.strptime(value, '%Y-%m-%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "datestrs = ['7/6/2011', '8/6/2011']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[datetime.datetime(2011, 7, 6, 0, 0), datetime.datetime(2011, 8, 6, 0, 0)]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[datetime.strptime(x, '%m/%d/%Y') for x in datestrs]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "对于一个一直的时间格式，使用datetime.strptime来解析日期是很好的方法。但是，如果每次都要写格式的话很烦人，尤其是对于一些比较常见的格式。在这种情况下，我们可以使用第三方库dateutil中的parser.parse方法（这个库会在安装pandas的时候自动安装）："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from dateutil.parser import parse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2011, 1, 3, 0, 0)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parse('2011-01-03')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "dateutil能够解析很多常见的时间表示格式："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(1997, 1, 31, 22, 45)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parse('Jan 31, 1997 10:45 PM')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在国际上，日在月之前是很常见的（译者：美国是把月放在日前面的），所以我们可以设置dayfirst=True来指明最前面的是否是日："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2011, 12, 6, 0, 0)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parse('6/12/2011', dayfirst=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "pandas通常可以用于处理由日期组成的数组，不论是否是DataFrame中的行索引或列。to_datetime方法能解析很多不同种类的日期表示。标准的日期格式，比如ISO 8601，能被快速解析："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "datestrs = ['2011-07-06 12:00:00', '2011-08-06 00:00:00']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['2011-07-06 12:00:00', '2011-08-06 00:00:00'], dtype='datetime64[ns]', freq=None)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.to_datetime(datestrs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "还能处理一些应该被判断为缺失的值（比如None, 空字符串之类的）："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['2011-07-06 12:00:00', '2011-08-06 00:00:00', 'NaT'], dtype='datetime64[ns]', freq=None)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "idx = pd.to_datetime(datestrs + [None])\n",
    "idx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "NaT"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "idx[2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False, False,  True], dtype=bool)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.isnull(idx)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nat(Not a Time)在pandas中，用于表示时间戳为空值（null value）。\n",
    "\n",
    "> dateutil.parse是一个很有用但不完美的工具。它可能会把一些字符串识别为日期，例如，'42'就会被解析为2042年加上今天的日期。\n",
    "\n",
    "datetime对象还有一些关于地区格式（locale-specific formatting）的选项，用于处理不同国家或不同语言的问题。例如，月份的缩写在德国和法国，与英语是不同的。下表列出一些相关的选项：\n",
    "\n",
    "![](http://oydgk2hgw.bkt.clouddn.com/pydata-book/gp2fy.png)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [py35]",
   "language": "python",
   "name": "Python [py35]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}