原文是 pandas documentation 中的 10 Minutes
to pandas
十分钟你妹啊！！
导入 pandas、numpy、matplotlib
12345In [1]: import pandas as pdIn [2]: import numpy as npIn [3]: import matplotlib.pyplot as plt
创造对象
Series
是一个值的序列，它只有一个列，以及索引。下面的例子中，就用默认的整数索引
1234567891011In [4]: s = pd.Series([1,3,5,np.nan,6,8])In [5]: sOut[5]: 0 11 32 53 NaN4 65 8dtype: float64
DataFrame 是有多个列的数据表，每个列拥有一个 label，当然，DataFrame
也有索引
12345678910111213141516171819In [6]: dates = pd.date_range('20130101', periods=6)In [7]: datesOut[7]: DatetimeIndex(['2013-01-01', '2013-01-02', '2013-01-03', '2013-01-04', '2013-01-05', '2013-01-06'], dtype='datetime64[ns]', freq='D')In [8]: df = pd.DataFrame(np.random.randn(6,4), index=dates, columns=list('ABCD'))In [9]: dfOut[9]: A B C D2013-01-01 0.469112 -0.282863 -1.509059 -1.1356322013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-03 -0.861849 -2.104569 -0.494929 1.0718042013-01-04 0.721555 -0.706771 -1.039575 0.2718602013-01-05 -0.424972 0.567020 0.276232 -1.0874012013-01-06 -0.673690 0.113648 -1.478427 0.524988
如果参数是一个 dict，每个 dict 的 value 会被转化成一个 Series
123456789101112131415In [10]: df2 = pd.DataFrame({ 'A' : 1., ....: 'B' : pd.Timestamp('20130102'), ....: 'C' : pd.Series(1,index=list(range(4)),dtype='float32'), ....: 'D' : np.array([3] * 4,dtype='int32'), ....: 'E' : pd.Categorical(["test","train","test","train"]), ....: 'F' : 'foo' }) ....: In [11]: df2Out[11]: A B C D E F0 1 2013-01-02 1 3 test foo1 1 2013-01-02 1 3 train foo2 1 2013-01-02 1 3 test foo3 1 2013-01-02 1 3 train foo
每列的格式用 dtypes 查看
123456789In [12]: df2.dtypesOut[12]: A float64B datetime64[ns]C float32D int32E categoryF objectdtype: object
你可以认为，DataFrame 是由 Series 组成的
1234567In [13]: df2.AOut[13]: 0 11 12 13 1Name: A, dtype: float64
查看数据
用 head 和 tail 查看顶端和底端的几列
123456789101112131415In [14]: df.head()Out[14]: A B C D2013-01-01 0.469112 -0.282863 -1.509059 -1.1356322013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-03 -0.861849 -2.104569 -0.494929 1.0718042013-01-04 0.721555 -0.706771 -1.039575 0.2718602013-01-05 -0.424972 0.567020 0.276232 -1.087401In [15]: df.tail(3)Out[15]: A B C D2013-01-04 0.721555 -0.706771 -1.039575 0.2718602013-01-05 -0.424972 0.567020 0.276232 -1.0874012013-01-06 -0.673690 0.113648 -1.478427 0.524988
实际上，DataFrame 内部用 numpy 格式存储数据。你也可以单独查看 index
和 columns
1234567891011121314151617In [16]: df.indexOut[16]: DatetimeIndex(['2013-01-01', '2013-01-02', '2013-01-03', '2013-01-04', '2013-01-05', '2013-01-06'], dtype='datetime64[ns]', freq='D')In [17]: df.columnsOut[17]: Index(['A', 'B', 'C', 'D'], dtype='object')In [18]: df.valuesOut[18]: array([[ 0.4691, -0.2829, -1.5091, -1.1356], [ 1.2121, -0.1732, 0.1192, -1.0442], [-0.8618, -2.1046, -0.4949, 1.0718], [ 0.7216, -0.7068, -1.0396, 0.2719], [-0.425 , 0.567 , 0.2762, -1.0874], [-0.6737, 0.1136, -1.4784, 0.525 ]])
describe() 显示数据的概要。
1234567891011In [19]: df.describe()Out[19]: A B C Dcount 6.000000 6.000000 6.000000 6.000000mean 0.073711 -0.431125 -0.687758 -0.233103std 0.843157 0.922818 0.779887 0.973118min -0.861849 -2.104569 -1.509059 -1.13563225% -0.611510 -0.600794 -1.368714 -1.07661050% 0.022070 -0.228039 -0.767252 -0.38618875% 0.658444 0.041933 -0.034326 0.461706max 1.212112 0.567020 0.276232 1.071804
和 numpy 一样，可以方便的得到转置
1234567In [20]: df.TOut[20]: 2013-01-01 2013-01-02 2013-01-03 2013-01-04 2013-01-05 2013-01-06A 0.469112 1.212112 -0.861849 0.721555 -0.424972 -0.673690B -0.282863 -0.173215 -2.104569 -0.706771 0.567020 0.113648C -1.509059 0.119209 -0.494929 -1.039575 0.276232 -1.478427D -1.135632 -1.044236 1.071804 0.271860 -1.087401 0.524988
对 axis 按照 index 排序（axis=1
是指第二个维度，即：列）
123456789In [21]: df.sort_index(axis=1, ascending=False)Out[21]: D C B A2013-01-01 -1.135632 -1.509059 -0.282863 0.4691122013-01-02 -1.044236 0.119209 -0.173215 1.2121122013-01-03 1.071804 -0.494929 -2.104569 -0.8618492013-01-04 0.271860 -1.039575 -0.706771 0.7215552013-01-05 -1.087401 0.276232 0.567020 -0.4249722013-01-06 0.524988 -1.478427 0.113648 -0.673690
按值排序
123456789In [22]: df.sort_values(by='B')Out[22]: A B C D2013-01-03 -0.861849 -2.104569 -0.494929 1.0718042013-01-04 0.721555 -0.706771 -1.039575 0.2718602013-01-01 0.469112 -0.282863 -1.509059 -1.1356322013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-06 -0.673690 0.113648 -1.478427 0.5249882013-01-05 -0.424972 0.567020 0.276232 -1.087401
选择
注意，以下这些对交互式环境很友好，但是作为 production code
请用优化过的 .at, .iat, .loc,
.iloc 和 .ix
获取行/列
从 DataFrame 选择一个列，就得到了 Series
123456789In [23]: df['A']Out[23]: 2013-01-01 0.4691122013-01-02 1.2121122013-01-03 -0.8618492013-01-04 0.7215552013-01-05 -0.4249722013-01-06 -0.673690Freq: D, Name: A, dtype: float64
和 numpy 类似，这里也能用 []
12345678910111213In [24]: df[0:3]Out[24]: A B C D2013-01-01 0.469112 -0.282863 -1.509059 -1.1356322013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-03 -0.861849 -2.104569 -0.494929 1.071804In [25]: df['20130102':'20130104']Out[25]: A B C D2013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-03 -0.861849 -2.104569 -0.494929 1.0718042013-01-04 0.721555 -0.706771 -1.039575 0.271860
通过 label 选择
刚刚那个 DataFrame
可以通过时间戳的下标（dates[0] = Timestamp('20130101')）来访问
1234567In [26]: df.loc[dates[0]]Out[26]: A 0.469112B -0.282863C -1.509059D -1.135632Name: 2013-01-01 00:00:00, dtype: float64
还可以多选
123456789In [27]: df.loc[:,['A','B']]Out[27]: A B2013-01-01 0.469112 -0.2828632013-01-02 1.212112 -0.1732152013-01-03 -0.861849 -2.1045692013-01-04 0.721555 -0.7067712013-01-05 -0.424972 0.5670202013-01-06 -0.673690 0.113648
注意那个冒号，用法和 MATLAB 或 NumPy 是一样的！所以也可以这样
123456In [28]: df.loc['20130102':'20130104',['A','B']]Out[28]: A B2013-01-02 1.212112 -0.1732152013-01-03 -0.861849 -2.1045692013-01-04 0.721555 -0.706771
依旧和 MATLAB
一样，当有一个维度是标量（而不是范围或序列）的时候，选择出的矩阵维度会减少
12345In [29]: df.loc['20130102',['A','B']]Out[29]: A 1.212112B -0.173215Name: 2013-01-02 00:00:00, dtype: float64
如果对所有的维度都写了标量，不就是选出一个元素吗？
12In [30]: df.loc[dates[0],'A']Out[30]: 0.46911229990718628
这种情况通常用 at ，速度更快
12In [31]: df.at[dates[0],'A']Out[31]: 0.46911229990718628
通过整数下标选择
和 MATLAB 完全一样
这个就和数组类似啦，直接看例子。选出第3行：
1234567In [32]: df.iloc[3]Out[32]: A 0.721555B -0.706771C -1.039575D 0.271860Name: 2013-01-04 00:00:00, dtype: float64
选出34行，01列：
12345In [33]: df.iloc[3:5,0:2]Out[33]: A B2013-01-04 0.721555 -0.7067712013-01-05 -0.424972 0.567020
也能用 list 选择
123456In [34]: df.iloc[[1,2,4],[0,2]]Out[34]: A C2013-01-02 1.212112 0.1192092013-01-03 -0.861849 -0.4949292013-01-05 -0.424972 0.276232
也能用 slice
12345In [35]: df.iloc[1:3,:]Out[35]: A B C D2013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-03 -0.861849 -2.104569 -0.494929 1.071804
123456789In [36]: df.iloc[:,1:3]Out[36]: B C2013-01-01 -0.282863 -1.5090592013-01-02 -0.173215 0.1192092013-01-03 -2.104569 -0.4949292013-01-04 -0.706771 -1.0395752013-01-05 0.567020 0.2762322013-01-06 0.113648 -1.478427
对应单个元素
12In [37]: df.iloc[1,1]Out[37]: -0.17321464905330858
12In [38]: df.iat[1,1]Out[38]: -0.17321464905330858
布尔值下标
和 MATLAB 类似
基本用法
123456In [39]: df[df.A > 0]Out[39]: A B C D2013-01-01 0.469112 -0.282863 -1.509059 -1.1356322013-01-02 1.212112 -0.173215 0.119209 -1.0442362013-01-04 0.721555 -0.706771 -1.039575 0.271860
没有填充的值等于 NaN
123456789In [40]: df[df > 0]Out[40]: A B C D2013-01-01 0.469112 NaN NaN NaN2013-01-02 1.212112 NaN 0.119209 NaN2013-01-03 NaN NaN NaN 1.0718042013-01-04 0.721555 NaN NaN 0.2718602013-01-05 NaN 0.567020 0.276232 NaN2013-01-06 NaN 0.113648 NaN 0.524988
isin() 函数：是否在集合中
12345678910111213141516171819In [41]: df2 = df.copy()In [42]: df2['E'] = ['one', 'one','two','three','four','three']In [43]: df2Out[43]: A B C D E2013-01-01 0.469112 -0.282863 -1.509059 -1.135632 one2013-01-02 1.212112 -0.173215 0.119209 -1.044236 one2013-01-03 -0.861849 -2.104569 -0.494929 1.071804 two2013-01-04 0.721555 -0.706771 -1.039575 0.271860 three2013-01-05 -0.424972 0.567020 0.276232 -1.087401 four2013-01-06 -0.673690 0.113648 -1.478427 0.524988 threeIn [44]: df2[df2['E'].isin(['two','four'])]Out[44]: A B C D E2013-01-03 -0.861849 -2.104569 -0.494929 1.071804 two2013-01-05 -0.424972 0.567020 0.276232 -1.087401 four
Setting
为 DataFrame 增加新的列，按 index 对应
12345678910111213In [45]: s1 = pd.Series([1,2,3,4,5,6], index=pd.date_range('20130102', periods=6))In [46]: s1Out[46]: 2013-01-02 12013-01-03 22013-01-04 32013-01-05 42013-01-06 52013-01-07 6Freq: D, dtype: int64In [47]: df['F'] = s1
通过 label 设置
1In [48]: df.at[dates[0],'A'] = 0
通过下标设置
1In [49]: df.iat[0,1] = 0
用 numpy 数组设置
1df.loc[:,'D'] = np.array([5] * len(df))
用布尔值作下标的 set
1In [53]: df2[df2 > 0] = -df2
缺失值
pandas 用 np.nan 表示缺失值。通常它不会被计算。
Reindexing 允许你改变某个轴的 index（以下代码制造一个示例用的
DataFrame）
1234567891011In [55]: df1 = df.reindex(index=dates[0:4], columns=list(df.columns) + ['E'])In [56]: df1.loc[dates[0]:dates[1],'E'] = 1In [57]: df1Out[57]: A B C D F E2013-01-01 0.000000 0.000000 -1.509059 5 NaN 12013-01-02 1.212112 -0.173215 0.119209 5 1 12013-01-03 -0.861849 -2.104569 -0.494929 5 2 NaN2013-01-04 0.721555 -0.706771 -1.039575 5 3 NaN
丢弃有 NaN 的行
1234In [58]: df1.dropna()Out[58]: A B C D F E2013-01-02 1.212112 -0.173215 0.119209 5 1 1
填充缺失值
1234567In [59]: df1.fillna(value=5)Out[59]: A B C D F E2013-01-01 0.000000 0.000000 -1.509059 5 5 12013-01-02 1.212112 -0.173215 0.119209 5 1 12013-01-03 -0.861849 -2.104569 -0.494929 5 2 52013-01-04 0.721555 -0.706771 -1.039575 5 3 5
获取布尔值的 mask：哪些值是 NaN
1234567In [60]: pd.isnull(df1)Out[60]: A B C D F E2013-01-01 False False False False True False2013-01-02 False False False False False False2013-01-03 False False False False False True2013-01-04 False False False False False True
操作
统计
通常，操作都会把 NaN 排除在外
平均值
12345678In [61]: df.mean()Out[61]: A -0.004474B -0.383981C -0.687758D 5.000000F 3.000000dtype: float64
对另一个维度做平均值，只要加个参数
123456789In [62]: df.mean(1)Out[62]: 2013-01-01 0.8727352013-01-02 1.4316212013-01-03 0.7077312013-01-04 1.3950422013-01-05 1.8836562013-01-06 1.592306Freq: D, dtype: float64
Apply
对数据（行或列） Apply 函数
123456789101112131415161718In [66]: df.apply(np.cumsum)Out[66]: A B C D F2013-01-01 0.000000 0.000000 -1.509059 5 NaN2013-01-02 1.212112 -0.173215 -1.389850 10 12013-01-03 0.350263 -2.277784 -1.884779 15 32013-01-04 1.071818 -2.984555 -2.924354 20 62013-01-05 0.646846 -2.417535 -2.648122 25 102013-01-06 -0.026844 -2.303886 -4.126549 30 15In [67]: df.apply(lambda x: x.max() - x.min())Out[67]: A 2.073961B 2.671590C 1.785291D 0.000000F 4.000000dtype: float64
直方图
1234567891011121314151617181920212223In [68]: s = pd.Series(np.random.randint(0, 7, size=10))In [69]: sOut[69]: 0 41 22 13 24 65 46 47 68 49 4dtype: int32In [70]: s.value_counts()Out[70]: 4 56 22 21 1dtype: int64
字符串函数
Series 自带了很多字符串处理函数，在 str
属性中，下面是一个例子
1234567891011121314In [71]: s = pd.Series(['A', 'B', 'C', 'Aaba', 'Baca', np.nan, 'CABA', 'dog', 'cat'])In [72]: s.str.lower()Out[72]: 0 a1 b2 c3 aaba4 baca5 NaN6 caba7 dog8 catdtype: object
Merge
Concat
简单地按行拼接
1234567891011121314151617181920212223242526272829303132In [73]: df = pd.DataFrame(np.random.randn(10, 4))In [74]: dfOut[74]: 0 1 2 30 -0.582002 0.066403 0.917236 -0.2141551 2.063923 1.930796 0.139574 0.4493682 -1.348962 0.228120 0.323906 1.2807783 0.689536 -0.083717 1.436075 0.6632504 -1.895829 -0.726235 -0.770739 0.1924825 0.302074 0.228735 1.390550 0.1961596 0.672059 -1.576747 0.154820 1.2188927 2.378061 0.280385 1.055607 -0.4692258 -0.997102 -0.533977 0.311215 0.9405709 -1.381892 -1.450002 0.562337 -1.195926# break it into piecesIn [75]: pieces = [df[3:7], df[:3], df[7:]]In [76]: pd.concat(pieces)Out[76]: 0 1 2 33 0.689536 -0.083717 1.436075 0.6632504 -1.895829 -0.726235 -0.770739 0.1924825 0.302074 0.228735 1.390550 0.1961596 0.672059 -1.576747 0.154820 1.2188920 -0.582002 0.066403 0.917236 -0.2141551 2.063923 1.930796 0.139574 0.4493682 -1.348962 0.228120 0.323906 1.2807787 2.378061 0.280385 1.055607 -0.4692258 -0.997102 -0.533977 0.311215 0.9405709 -1.381892 -1.450002 0.562337 -1.195926
Join
和 SQL 的 join 是一个意思
1234567891011121314151617181920212223In [77]: left = pd.DataFrame({'key': ['foo', 'foo'], 'lval': [1, 2]})In [78]: right = pd.DataFrame({'key': ['foo', 'foo'], 'rval': [4, 5]})In [79]: leftOut[79]: key lval0 foo 11 foo 2In [80]: rightOut[80]: key rval0 foo 41 foo 5In [81]: pd.merge(left, right, on='key')Out[81]: key lval rval0 foo 1 41 foo 1 52 foo 2 43 foo 2 5
Append
向 DataFrame 增加新的数据行
12345678910111213141516171819202122232425262728In [82]: df = pd.DataFrame(np.random.randn(8, 4), columns=['A','B','C','D'])In [83]: dfOut[83]: A B C D0 1.346061 1.511763 1.627081 -0.9905821 -0.441652 1.211526 0.268520 0.0245802 -1.577585 0.396823 -0.105381 -0.5325323 1.453749 1.208843 -0.080952 -0.2646104 -0.727965 -0.589346 0.339969 -0.6932055 -0.339355 0.593616 0.884345 1.5914316 0.141809 0.220390 0.435589 0.1924517 -0.096701 0.803351 1.715071 -0.708758In [84]: s = df.iloc[3]In [85]: df.append(s, ignore_index=True)Out[85]: A B C D0 1.346061 1.511763 1.627081 -0.9905821 -0.441652 1.211526 0.268520 0.0245802 -1.577585 0.396823 -0.105381 -0.5325323 1.453749 1.208843 -0.080952 -0.2646104 -0.727965 -0.589346 0.339969 -0.6932055 -0.339355 0.593616 0.884345 1.5914316 0.141809 0.220390 0.435589 0.1924517 -0.096701 0.803351 1.715071 -0.7087588 1.453749 1.208843 -0.080952 -0.264610
Grouping
和 SQL 中的 GROUP BY 类似，包括以下这几步：
根据某些规则，把数据分组
对每组应用一个聚集函数，把结果放在一个数据结构中
准备一下测试用的数据集
12345678910111213141516171819In [86]: df = pd.DataFrame({'A' : ['foo', 'bar', 'foo', 'bar', ....: 'foo', 'bar', 'foo', 'foo'], ....: 'B' : ['one', 'one', 'two', 'three', ....: 'two', 'two', 'one', 'three'], ....: 'C' : np.random.randn(8), ....: 'D' : np.random.randn(8)}) ....: In [87]: dfOut[87]: A B C D0 foo one -1.202872 -0.0552241 bar one -1.814470 2.3959852 foo two 1.018601 1.5528253 bar three -0.595447 0.1665994 foo two 1.395433 0.0476095 bar two -0.392670 -0.1364736 foo one 0.007207 -0.5617577 foo three 1.928123 -1.623033
做 Group 操作并对每组求和
123456In [88]: df.groupby('A').sum()Out[88]: C DA bar -2.802588 2.42611foo 3.146492 -0.63958
可以对两列进行 Group by 并求和
12345678910In [89]: df.groupby(['A','B']).sum()Out[89]: C DA B bar one -1.814470 2.395985 three -0.595447 0.166599 two -0.392670 -0.136473foo one -1.195665 -0.616981 three 1.928123 -1.623033 two 2.414034 1.600434
Reshape
Stack 层叠
准备一下数据
1234567891011121314151617181920In [90]: tuples = list(zip(*[['bar', 'bar', 'baz', 'baz', ....: 'foo', 'foo', 'qux', 'qux'], ....: ['one', 'two', 'one', 'two', ....: 'one', 'two', 'one', 'two']])) ....: In [91]: index = pd.MultiIndex.from_tuples(tuples, names=['first', 'second'])In [92]: df = pd.DataFrame(np.random.randn(8, 2), index=index, columns=['A', 'B'])In [93]: df2 = df[:4]In [94]: df2Out[94]: A Bfirst second bar one 0.029399 -0.542108 two 0.282696 -0.087302baz one -1.575170 1.771208 two 0.816482 1.100230
stack() 把 DataFrame 的列“压缩”到 index 里去
1234567891011121314In [95]: stacked = df2.stack()In [96]: stackedOut[96]: first second bar one A 0.029399 B -0.542108 two A 0.282696 B -0.087302baz one A -1.575170 B 1.771208 two A 0.816482 B 1.100230dtype: float64
反之，只要是 MultiIndex 都可以用 unstack()
恢复出列，默认把最后一个 index 解开
1234567891011121314151617181920212223242526In [97]: stacked.unstack()Out[97]: A Bfirst second bar one 0.029399 -0.542108 two 0.282696 -0.087302baz one -1.575170 1.771208 two 0.816482 1.100230In [98]: stacked.unstack(1)Out[98]: second one twofirst bar A 0.029399 0.282696 B -0.542108 -0.087302baz A -1.575170 0.816482 B 1.771208 1.100230In [99]: stacked.unstack(0)Out[99]: first bar bazsecond one A 0.029399 -1.575170 B -0.542108 1.771208two A 0.282696 0.816482 B -0.087302 1.100230
Pivot Table 旋转
准备一下数据
12345678910111213141516171819202122In [100]: df = pd.DataFrame({'A' : ['one', 'one', 'two', 'three'] * 3, .....: 'B' : ['A', 'B', 'C'] * 4, .....: 'C' : ['foo', 'foo', 'foo', 'bar', 'bar', 'bar'] * 2, .....: 'D' : np.random.randn(12), .....: 'E' : np.random.randn(12)}) .....: In [101]: dfOut[101]: A B C D E0 one A foo 1.418757 -0.1796661 one B foo -1.879024 1.2918362 two C foo 0.536826 -0.0096143 three A bar 1.006160 0.3921494 one B bar -0.029716 0.2645995 one C bar -1.146178 -0.0574096 two A foo 0.100900 -1.4256387 three B foo -1.035018 1.0240988 one C foo 0.314665 -0.1060629 one A bar -0.773723 1.82437510 two B bar -1.170653 0.59597411 three C bar 0.648740 1.167115
pivot 是把原来的数据(values)作为新表的行(index)、列(columns)
12345678910111213In [102]: pd.pivot_table(df, values='D', index=['A', 'B'], columns=['C'])Out[102]: C bar fooA B one A -0.773723 1.418757 B -0.029716 -1.879024 C -1.146178 0.314665three A 1.006160 NaN B NaN -1.035018 C 0.648740 NaNtwo A NaN 0.100900 B -1.170653 NaN C NaN 0.536826
时间序列
pandas 的时间序列功能在金融应用中很有用。
resample 功能：
123456789In [103]: rng = pd.date_range('1/1/2012', periods=100, freq='S')In [104]: ts = pd.Series(np.random.randint(0, 500, len(rng)), index=rng)In [105]: ts.resample('T', how='sum')Out[105]: 2012-01-01 00:00:00 148332012-01-01 00:01:00 9246Freq: T, dtype: int32
时区表示
1234567891011121314151617181920212223In [106]: rng = pd.date_range('3/6/2012 00:00', periods=5, freq='D')In [107]: ts = pd.Series(np.random.randn(len(rng)), rng)In [108]: tsOut[108]: 2012-03-06 0.4640002012-03-07 0.2273712012-03-08 -0.4969222012-03-09 0.3063892012-03-10 -2.290613Freq: D, dtype: float64In [109]: ts_utc = ts.tz_localize('UTC')In [110]: ts_utcOut[110]: 2012-03-06 00:00:00+00:00 0.4640002012-03-07 00:00:00+00:00 0.2273712012-03-08 00:00:00+00:00 -0.4969222012-03-09 00:00:00+00:00 0.3063892012-03-10 00:00:00+00:00 -2.290613Freq: D, dtype: float64
时区的转换：
12345678In [111]: ts_utc.tz_convert('US/Eastern')Out[111]: 2012-03-05 19:00:00-05:00 0.4640002012-03-06 19:00:00-05:00 0.2273712012-03-07 19:00:00-05:00 -0.4969222012-03-08 19:00:00-05:00 0.3063892012-03-09 19:00:00-05:00 -2.290613Freq: D, dtype: float64
从 Timestamp index 转换成 TimePeriod
1234567891011121314151617181920212223242526272829303132In [112]: rng = pd.date_range('1/1/2012', periods=5, freq='M')In [113]: ts = pd.Series(np.random.randn(len(rng)), index=rng)In [114]: tsOut[114]: 2012-01-31 -1.1346232012-02-29 -1.5618192012-03-31 -0.2608382012-04-30 0.2819572012-05-31 1.523962Freq: M, dtype: float64In [115]: ps = ts.to_period()In [116]: psOut[116]: 2012-01 -1.1346232012-02 -1.5618192012-03 -0.2608382012-04 0.2819572012-05 1.523962Freq: M, dtype: float64In [117]: ps.to_timestamp()Out[117]: 2012-01-01 -1.1346232012-02-01 -1.5618192012-03-01 -0.2608382012-04-01 0.2819572012-05-01 1.523962Freq: MS, dtype: float64
类别
1In [122]: df = pd.DataFrame({"id":[1,2,3,4,5,6], "raw_grade":['a', 'b', 'b', 'a', 'a', 'e']})
把上述的文本类型的 raw_grade 转化成类别：
123456789101112In [123]: df["grade"] = df["raw_grade"].astype("category")In [124]: df["grade"]Out[124]: 0 a1 b2 b3 a4 a5 eName: grade, dtype: categoryCategories (3, object): [a, b, e]
类别可以 inplace 地赋值：（只是改一下对应的字符串嘛，类别是用 Index
对象存储的）
1In [125]: df["grade"].cat.categories = ["very good", "good", "very bad"]
修改类别时，如果有新的类别，会自动加进去
123456789101112In [126]: df["grade"] = df["grade"].cat.set_categories(["very bad", "bad", "medium", "good", "very good"])In [127]: df["grade"]Out[127]: 0 very good1 good2 good3 very good4 very good5 very badName: grade, dtype: categoryCategories (5, object): [very bad, bad, medium, good, very good]
做 group by 的时候，空的类别也会被呈现出来
123456789In [129]: df.groupby("grade").size()Out[129]: gradevery bad 1bad 0medium 0good 2very good 3dtype: int64
对于 DataFrame，可以直接 plot
12345678In [133]: df = pd.DataFrame(np.random.randn(1000, 4), index=ts.index, .....: columns=['A', 'B', 'C', 'D']) .....: In [134]: df = df.cumsum()In [135]: plt.figure(); df.plot(); plt.legend(loc='best')Out[135]:
读取、写入数据
CSV
写入
1In [136]: df.to_csv('foo.csv')
读取
1234567891011121314151617181920In [137]: pd.read_csv('foo.csv')Out[137]: Unnamed: 0 A B C D0 2000-01-01 0.266457 -0.399641 -0.219582 1.1868601 2000-01-02 -1.170732 -0.345873 1.653061 -0.2829532 2000-01-03 -1.734933 0.530468 2.060811 -0.5155363 2000-01-04 -1.555121 1.452620 0.239859 -1.1568964 2000-01-05 0.578117 0.511371 0.103552 -2.4282025 2000-01-06 0.478344 0.449933 -0.741620 -1.9624096 2000-01-07 1.235339 -0.091757 -1.543861 -1.084753.. ... ... ... ... ...993 2002-09-20 -10.628548 -9.153563 -7.883146 28.313940994 2002-09-21 -10.390377 -8.727491 -6.399645 30.914107995 2002-09-22 -8.985362 -8.485624 -4.669462 31.367740996 2002-09-23 -9.558560 -8.781216 -4.499815 30.518439997 2002-09-24 -9.902058 -9.340490 -4.386639 30.105593998 2002-09-25 -10.216020 -9.480682 -3.933802 29.758560999 2002-09-26 -11.856774 -10.671012 -3.216025 29.369368[1000 rows x 5 columns]
HDF5
123In [138]: df.to_hdf('foo.h5','df')In [139]: pd.read_hdf('foo.h5','df')
Excel
123In [140]: df.to_excel('foo.xlsx', sheet_name='Sheet1')In [141]: pd.read_excel('foo.xlsx', 'Sheet1', index_col=None, na_values=['NA'])
坑
12345>>> if pd.Series([False, True, False]): print("I was true")Traceback ...ValueError: The truth value of an array is ambiguous. Use a.empty, a.any() or a.all().
如上，不能直接把返回值当作布尔值。
