介绍操作Series和DataFrame中的数据的基本功能
重新索引
pandas对象的一个重要方法是reindex,其作用是创建一个适应新索引的新对象。以之前的一个简单示例来说
In [1]: from pandas import Series,DataFrame
In [2]: import pandas as pd
In [3]: import numpy as np
In [4]: obj=Series([6.5,7.8,-5.9,8.6],index=['d','b','a','c'])
In [5]: obj
Out[5]:
d 6.5
b 7.8
a -5.9
c 8.6
dtype: float64
调用该Series的reindex将会根据新索引进行重排。如果某个索引值当前不存在,就引入缺失值
In [6]: obj2=obj.reindex(['a', 'b', 'c', 'd', 'e'])
In [7]: obj2
Out[7]:
a -5.9
b 7.8
c 8.6
d 6.5
e NaN
dtype: float64
In [8]: obj.reindex(['a', 'b', 'c', 'd', 'e'], fill_value=0)
Out[8]:
a -5.9
b 7.8
c 8.6
d 6.5
e 0.0
dtype: float64
In [9]: obj3=Series(['blue', 'purple', 'yellow'], index=[0, 2, 4])
In [10]: obj3.reindex(range(6), method='ffill')
Out[10]:
0 blue
1 blue
2 purple
3 purple
4 yellow
5 yellow
dtype: object
In [11]: obj3.reindex(range(6), method='bfill')
Out[11]:
0 blue
1 purple
2 purple
3 yellow
4 yellow
5 NaN
dtype: object
In [12]: obj3.reindex(range(6), method='pad')
Out[12]:
0 blue
1 blue
2 purple
3 purple
4 yellow
5 yellow
dtype: object
对于DataFrame,reindex可以修改(行)索引、列,或两个都修改。如果仅传入一个序列,则会重新索引行
In [13]: frame = DataFrame(np.arange(9).reshape((3, 3)), index=['a', 'c', 'd'],columns=['Ohio', 'Texas', 'California'])
In [14]: frame
Out[14]:
Ohio Texas California
a 0 1 2
c 3 4 5
d 6 7 8
In [15]: frame2=frame.reindex(['a', 'b', 'c', 'd'])
In [16]: frame2
Out[16]:
Ohio Texas California
a 0.0 1.0 2.0
b NaN NaN NaN
c 3.0 4.0 5.0
d 6.0 7.0 8.0
使用columns关键字即可重新索引列
In [17]: states = ['Texas', 'Utah', 'California']
In [18]: frame.reindex(columns=states)
Out[18]:
Texas Utah California
a 1 NaN 2
c 4 NaN 5
d 7 NaN 8
利用ix的标签索引功能
In [28]: frame
Out[28]:
Ohio Texas California
a 0 1 2
c 3 4 5
d 6 7 8
In [31]: states = ['Texas', 'Utah', 'California']
In [32]: frame.ix[['a', 'b', 'c', 'd'], states]
Out[32]:
Texas Utah California
a 1.0 NaN 2.0
b NaN NaN NaN
c 4.0 NaN 5.0
d 7.0 NaN 8.0
丢弃某条轴上的一个或多个项很简单,只要有一个索引数组或列表即可。由于需要执行一些数据整理和集合逻辑,所以drop方法返回的是一个在指定轴上删除了指定值的新对象
In [33]: obj=Series(np.arange(5.),index=['a', 'b', 'c', 'd', 'e'])
In [34]: obj
Out[34]:
a 0.0
b 1.0
c 2.0
d 3.0
e 4.0
dtype: float64
In [35]: new_obj=obj.drop('c')
In [36]: new_obj
Out[36]:
a 0.0
b 1.0
d 3.0
e 4.0
dtype: float64
In [37]: obj.drop(['d','b'])
Out[37]:
a 0.0
c 2.0
e 4.0
dtype: float64
对于DataFrame,可以删除任意轴上的索引值
In [41]: data = DataFrame(np.arange(16).reshape((4, 4)),
...: index=['Ohio', 'Colorado', 'Utah', 'New York'],
...: columns=['one', 'two', 'three', 'four'])
In [42]: data
Out[42]:
one two three four
Ohio 0 1 2 3
Colorado 4 5 6 7
Utah 8 9 10 11
New York 12 13 14 15
In [43]: data.drop(['Colorado', 'Ohio'])
Out[43]:
one two three four
Utah 8 9 10 11
New York 12 13 14 15
In [44]: data.drop('two',axis=1)
Out[44]:
one three four
Ohio 0 2 3
Colorado 4 6 7
Utah 8 10 11
New York 12 14 15
In [45]: data.drop(['two', 'four'], axis=1)
Out[45]:
one three
Ohio 0 2
Colorado 4 6
Utah 8 10
New York 12 14
索引、选取和过滤
Series的索引值不只是整数
In [47]: obj=Series(np.arange(5.),index=['a', 'b', 'c', 'd','e'])
In [48]: obj
Out[48]:
a 0.0
b 1.0
c 2.0
d 3.0
e 4.0
dtype: float64
In [49]: obj['b']
Out[49]: 1.0
In [50]: obj[3]
Out[50]: 3.0
In [51]: obj[3:5]
Out[51]:
d 3.0
e 4.0
dtype: float64
In [52]: obj[['b','e','d']]
Out[52]:
b 1.0
e 4.0
d 3.0
dtype: float64
In [53]: obj[[1,4]]
Out[53]:
b 1.0
e 4.0
dtype: float64
In [54]: obj[obj<3]
Out[54]:
a 0.0
b 1.0
c 2.0
dtype: float64
利用标签的切片运算与普通的Python切片运算不同,其末端是包含的(inclusive)
In [55]: obj['b':'d']
Out[55]:
b 1.0
c 2.0
d 3.0
dtype: float64
In [56]: obj['b':'d']=6
In [57]: obj
Out[57]:
a 0.0
b 6.0
c 6.0
d 6.0
e 4.0
dtype: float64
DataFrame进行索引其实就是获取一个或多个列
In [60]: data = DataFrame(np.arange(16).reshape((4, 4)),
...: index=['Ohio', 'Colorado', 'Utah', 'New York'],
...: columns=['one', 'two', 'three', 'four'])
In [61]: data
Out[61]:
one two three four
Ohio 0 1 2 3
Colorado 4 5 6 7
Utah 8 9 10 11
New York 12 13 14 15
In [62]: data['two']
Out[62]:
Ohio 1
Colorado 5
Utah 9
New York 13
Name: two, dtype: int32
In [63]: data[['three','one']]
Out[63]:
three one
Ohio 2 0
Colorado 6 4
Utah 10 8
New York 14 12
通过切片或布尔型数组选取行
In [64]: data[:3]
Out[64]:
one two three four
Ohio 0 1 2 3
Colorado 4 5 6 7
Utah 8 9 10 11
In [65]: data[data['three']>5]
Out[65]:
one two three four
Colorado 4 5 6 7
Utah 8 9 10 11
New York 12 13 14 15
通过布尔型DataFrame(比如下面由标量比较运算得出的)进行索引
In [66]: data<6
Out[66]:
one two three four
Ohio True True True True
Colorado True True False False
Utah False False False False
New York False False False False
In [67]: data[data<5]=0
In [68]: data
Out[68]:
one two three four
Ohio 0 0 0 0
Colorado 0 5 6 7
Utah 8 9 10 11
New York 12 13 14 15
利用索引字段ix,它可以通过NumPy式的标记法以及轴标签从DataFrame中选取行和列的子集。其中:ix is deprecated,可以使用loc
In [69]: data.ix['Colorado', ['two', 'three']]
C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:1: DeprecationWarning:
.ix is deprecated. Please use
.loc for label based indexing or
.iloc for positional indexing
See the documentation here:
http://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate_ix
"""Entry point for launching an IPython kernel.
Out[69]:
two 5
three 6
Name: Colorado, dtype: int32
In [70]: data.loc['Colorado', ['two', 'three']]
Out[70]:
two 5
three 6
Name: Colorado, dtype: int32
In [71]: data.ix[['Colorado', 'Utah'], [3, 0, 1]]
Out[71]:
four one two
Colorado 7 0 5
Utah 11 8 9
In [72]: data.ix[2]
Out[72]:
one 8
two 9
three 10
four 11
Name: Utah, dtype: int32
In [73]: data.loc[:'Utah','two']
Out[73]:
Ohio 0
Colorado 5
Utah 9
Name: two, dtype: int32
In [74]: data.ix[data.three>5]
Out[74]:
one two three four
Colorado 0 5 6 7
Utah 8 9 10 11
New York 12 13 14 15
In [75]: data.ix[data.three > 5, :3]
Out[75]:
one two three
Colorado 0 5 6
Utah 8 9 10
New York 12 13 14
对pandas对象中的数据的选取和重排方式有很多
为什么不是输出7 4 5,而输出的是7 0 5,是不能理解的小地方,只能慢慢体会其中的用法。
其中,get_value方法是选取,set-value方法是设置
算术运算和数据对齐
Pandas可以对不同索引的对象进行算术运算。在将对象相加时,如果存在不同的索引对,则结果的索引就是该索引对的并集。
In [77]: s1=Series([6.8,-4.5,3.6,5.6],index=['a','c','d','e'])
In [78]: s2 = Series([-6.5, 3.6, -5.6, 4, 3.1], index=['a', 'c', 'e', 'f', 'g'])
In [79]: s1
Out[79]:
a 6.8
c -4.5
d 3.6
e 5.6
dtype: float64
In [80]: s2
Out[80]:
a -6.5
c 3.6
e -5.6
f 4.0
g 3.1
dtype: float64
In [81]: s1+s2
Out[81]:
a 0.3
c -0.9
d NaN
e 0.0
f NaN
g NaN
dtype: float64
自动的数据对齐操作在不重叠的索引处引入了NA值。缺失值会在算术运算过程中传播。对于DataFrame,对齐操作会同时发生在行和列上。相加后将会返回一个新的DataFrame,其索引和列为原来那两个DataFrame的并集
In [85]: df1 = DataFrame(np.arange(9.).reshape((3, 3)), columns=list('bcd'),
...: index=['Ohio', 'Texas', 'Colorado'])
In [86]: df2 = DataFrame(np.arange(12.).reshape((4, 3)), columns=list('bde'),
...: index=['Utah', 'Ohio', 'Texas', 'Oregon'])
In [87]: df1
Out[87]:
b c d
Ohio 0.0 1.0 2.0
Texas 3.0 4.0 5.0
Colorado 6.0 7.0 8.0
In [88]: df2
Out[88]:
b d e
Utah 0.0 1.0 2.0
Ohio 3.0 4.0 5.0
Texas 6.0 7.0 8.0
Oregon 9.0 10.0 11.0
In [89]: df1+df2
Out[89]:
b c d e
Colorado NaN NaN NaN NaN
Ohio 3.0 NaN 6.0 NaN
Oregon NaN NaN NaN NaN
Texas 9.0 NaN 12.0 NaN
Utah NaN NaN NaN NaN
在算术方法中填充值
在对不同索引的对象进行算术运算时,你可能希望当一个对象中某个轴标签在另一个对象中找不到时填充一个特殊值(比如0),相加时,没有重叠的位置就会产生NA值。
In [95]: f1 = DataFrame(np.arange(12.).reshape((3, 4)), columns=list('abcd'))
In [96]: f2 = DataFrame(np.arange(20.).reshape((4, 5)), columns=list('abcde'))
In [97]: f1
Out[97]:
a b c d
0 0.0 1.0 2.0 3.0
1 4.0 5.0 6.0 7.0
2 8.0 9.0 10.0 11.0
In [98]: f2
Out[98]:
a b c d e
0 0.0 1.0 2.0 3.0 4.0
1 5.0 6.0 7.0 8.0 9.0
2 10.0 11.0 12.0 13.0 14.0
3 15.0 16.0 17.0 18.0 19.0
In [99]: f1+f2
Out[99]:
a b c d e
0 0.0 2.0 4.0 6.0 NaN
1 9.0 11.0 13.0 15.0 NaN
2 18.0 20.0 22.0 24.0 NaN
3 NaN NaN NaN NaN NaN
使用add方法,传入f2以及一个fill_value参数
In [102]: f1.add(f2, fill_value=0)
Out[102]:
a b c d e
0 0.0 2.0 4.0 6.0 4.0
1 9.0 11.0 13.0 15.0 9.0
2 18.0 20.0 22.0 24.0 14.0
3 15.0 16.0 17.0 18.0 19.0
在对Series或DataFrame重新索引时,也可以指定一个填充值
In [103]: f1.reindex(columns=f2.columns, fill_value=0)
Out[103]:
a b c d e
0 0.0 1.0 2.0 3.0 0
1 4.0 5.0 6.0 7.0 0
2 8.0 9.0 10.0 11.0 0
In [105]: f1*f2
Out[105]:
a b c d e
0 0.0 1.0 4.0 9.0 NaN
1 20.0 30.0 42.0 56.0 NaN
2 80.0 99.0 120.0 143.0 NaN
3 NaN NaN NaN NaN NaN
DataFrame和Series之间的运算
计算一个二维数组与其某行之间的差,出现的结果这就叫做广播(broadcasting),如下:
In [106]: arr=np.arange(12.).reshape((3,4))
In [107]: arr
Out[107]:
array([[ 0., 1., 2., 3.],
[ 4., 5., 6., 7.],
[ 8., 9., 10., 11.]])
In [108]: arr[0]
Out[108]: array([ 0., 1., 2., 3.])
In [109]: arr-arr[0]
Out[109]:
array([[ 0., 0., 0., 0.],
[ 4., 4., 4., 4.],
[ 8., 8., 8., 8.]])
默认情况下,DataFrame和Series之间的算术运算会将Series的索引匹配到DataFrame的列,然后沿着行一直向下广播
In [110]: frame = DataFrame(np.arange(12.).reshape((4, 3)), columns=list('bde'),
...: index=['Utah', 'Ohio', 'Texas', 'Oregon'])
In [111]: frame
Out[111]:
b d e
Utah 0.0 1.0 2.0
Ohio 3.0 4.0 5.0
Texas 6.0 7.0 8.0
Oregon 9.0 10.0 11.0
In [112]: series = frame.ix[0]
In [113]: series
Out[113]:
b 0.0
d 1.0
e 2.0
Name: Utah, dtype: float64
In [114]: frame-series
Out[114]:
b d e
Utah 0.0 0.0 0.0
Ohio 3.0 3.0 3.0
Texas 6.0 6.0 6.0
Oregon 9.0 9.0 9.0
如果某个索引值在DataFrame的列或Series的索引中找不到,则参与运算的两个对象就会被重新索引以形成并集
In [115]: series2 = Series(range(3), index=['b', 'e', 'f'])
In [116]: series2
Out[116]:
b 0
e 1
f 2
dtype: int32
In [117]: frame + series2
Out[117]:
b d e f
Utah 0.0 NaN 3.0 NaN
Ohio 3.0 NaN 6.0 NaN
Texas 6.0 NaN 9.0 NaN
Oregon 9.0 NaN 12.0 NaN
如果你希望匹配行且在列上广播,则必须使用算术运算方法。
In [118]: series3 = frame['d']
In [119]: series3
Out[119]:
Utah 1.0
Ohio 4.0
Texas 7.0
Oregon 10.0
Name: d, dtype: float64
In [120]: frame
Out[120]:
b d e
Utah 0.0 1.0 2.0
Ohio 3.0 4.0 5.0
Texas 6.0 7.0 8.0
Oregon 9.0 10.0 11.0
In [121]: frame.sub(series3, axis=0)
Out[121]:
b d e
Utah -1.0 0.0 1.0
Ohio -1.0 0.0 1.0
Texas -1.0 0.0 1.0
Oregon -1.0 0.0 1.0
传入的轴号就是希望匹配的轴。
函数应用和映射
NumPy的ufuncs(元素级数组方法)也可用于操作pandas对象
In [122]: frame = DataFrame(np.random.randn(4, 3), columns=list('bde'),
...: index=['Utah', 'Ohio', 'Texas', 'Oregon'])
In [123]: frame
Out[123]:
b d e
Utah -0.976531 -1.511940 -0.018721
Ohio 0.598117 0.047678 -0.058404
Texas 2.469704 0.027215 1.154004
Oregon 1.308615 -1.634739 0.096210
In [124]: np.abs(frame)
Out[124]:
b d e
Utah 0.976531 1.511940 0.018721
Ohio 0.598117 0.047678 0.058404
Texas 2.469704 0.027215 1.154004
Oregon 1.308615 1.634739 0.096210
将函数应用到由各列或行所形成的一维数组上。DataFrame的apply方法即可实现此功能。
In [125]: f = lambda x: x.max() - x.min()
In [126]: frame.apply(f)
Out[126]:
b 3.446234
d 1.682417
e 1.212408
dtype: float64
In [127]: frame.apply(f,axis=1)
Out[127]:
Utah 1.493219
Ohio 0.656521
Texas 2.442489
Oregon 2.943355
dtype: float64
sum和mean方法
In [128]: def f(x):
...: return Series([x.min(), x.max()], index=['min', 'max'])
...:
In [129]: frame.apply(f)
Out[129]:
b d e
min -0.976531 -1.634739 -0.058404
max 2.469704 0.047678 1.154004
得到frame中各个浮点值的格式化字符串,使用applymap
In [128]: def f(x):
...:
return Series([x.min(), x.max()], index=['min', 'max'])
In [129]: frame.apply(f)
Out[129]:
b d e
min -0.976531 -1.634739 -0.058404
max 2.469704 0.047678 1.154004
In [130]: format = lambda x: '%.2f' % x
In [131]: frame.applymap(format)
Out[131]:
b d e
Utah -0.98 -1.51 -0.02
Ohio 0.60 0.05 -0.06
Texas 2.47 0.03 1.15
Oregon 1.31 -1.63 0.10
Series有一个用于应用元素级函数的map方法
In [132]: frame['e'].map(format)
Out[132]:
Utah -0.02
Ohio -0.06
Texas 1.15
Oregon 0.10
Name: e, dtype: object
排序和排名
根据条件对数据集排序(sorting)也是一种重要的内置运算。要对行或列索引进行排序(按字典顺序),可使用sort_index方法,它将返回一个已排序的新对象
In [133]: obj = Series(range(4), index=['d', 'a', 'b', 'c'])
In [134]: obj
Out[134]:
d 0
a 1
b 2
c 3
dtype: int32
In [135]: obj.sort_index()
Out[135]:
a 1
b 2
c 3
d 0
dtype: int32
DataFrame,则可以根据任意一个轴上的索引进行排序
In [136]: frame = DataFrame(np.arange(8).reshape((2, 4)), index=['three', 'one'],
...: columns=['d', 'a', 'b', 'c'])
In [137]: frame
Out[137]:
d a b c
three 0 1 2 3
one 4 5 6 7
In [138]: frame.sort_index()
Out[138]:
d a b c
one 4 5 6 7
three 0 1 2 3
In [139]: frame.sort_index(axis=1)
Out[139]:
a b c d
three 1 2 3 0
one 5 6 7 4
数据默认是按升序排序的,但也可以降序排序
In [140]: frame.sort_index(axis=1,ascending=False)
Out[140]:
d c b a
three 0 3 2 1
one 4 7 6 5
series通过索引进行排序
In [148]: obj = Series([6, 9, -8, 3])
In [149]: obj.sort_index()
Out[149]:
0 6
1 9
2 -8
3 3
dtype: int64
series通过升值进行排序
In [150]: obj.sort_values()
Out[150]:
2 -8
3 3
0 6
1 9
dtype: int64
在排序时,任何缺失值默认都会被放到Series的末尾,其中order不能排序,使用sort_values进行排序
In [151]: obj = Series([4, np.nan, 6, np.nan, -3, 3])
In [152]: obj.order()
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-152-4fc888977b98> in <module>()
----> 1 obj.order()
C:\ProgramData\Anaconda3\lib\site-packages\pandas\core\generic.py in __getattr__(self, name)
2968 if name in self._info_axis:
2969 return self[name]
-> 2970 return object.__getattribute__(self, name)
2971
2972 def __setattr__(self, name, value):
AttributeError: 'Series' object has no attribute 'order'
In [153]: obj.sort_values()
Out[153]:
4 -3.0
5 3.0
0 4.0
2 6.0
1 NaN
3 NaN
dtype: float64
希望根据一个或多个列中的值进行排序,可以将一个或多个列的名字传递给by选项
In [154]: frame = DataFrame({'b': [5, 8, -6, 3], 'a': [0, 1, 0, 1]})
In [155]: frame
Out[155]:
a b
0 0 5
1 1 8
2 0 -6
3 1 3
In [156]: frame.sort_index(by='b')
Out[156]:
a b
2 0 -6
3 1 3
0 0 5
1 1 8
In [157]: frame.sort_index(by=['a','b'])
C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:1: FutureWarning: by argument to sort_index is deprecated, pls use .sort_values(by=...)
"""Entry point for launching an IPython kernel.
Out[157]:
a b
2 0 -6
0 0 5
3 1 3
1 1 8
要根据多个列进行排序,传入名称的列表
In [158]: frame.sort_values(by=['a','b'])
Out[158]:
a b
2 0 -6
0 0 5
3 1 3
1 1 8
排名(ranking)跟排序关系密切,且它会增设一个排名值(从1开始,一直到数组中有效数据的数量)。它跟numpy.argsort产生的间接排序索引差不多,只不过它可以根据某种规则破坏平级关系。默认情况下,rank是通过“为各组分配一个平均排名”的方式破坏平级关系的:
In [159]: obj = Series([8, -6, 5, 4, 2, 0, 4])
In [160]: obj
Out[160]:
0 8
1 -6
2 5
3 4
4 2
5 0
6 4
dtype: int64
In [161]: obj.rank()
Out[161]:
0 7.0
1 1.0
2 6.0
3 4.5
4 3.0
5 2.0
6 4.5
dtype: float64
可以根据值在原数据中出现的顺序给出排名
In [162]: obj.rank(method='first')
Out[162]:
0 7.0
1 1.0
2 6.0
3 4.0
4 3.0
5 2.0
6 5.0
dtype: float64
按降序进行排名
In [163]: obj.rank(ascending=False, method='max')
Out[163]:
0 1.0
1 7.0
2 2.0
3 4.0
4 5.0
5 6.0
6 4.0
dtype: float64
在行或列上计算排名
In [164]: frame = DataFrame({'b': [4.3, 7, -3, 2], 'a': [0, 1, 0, 1],
...: 'c': [-2, 5, 8, -2.5]})
In [165]: frame
Out[165]:
a b c
0 0 4.3 -2.0
1 1 7.0 5.0
2 0 -3.0 8.0
3 1 2.0 -2.5
In [166]: frame.rank(axis=1)
Out[166]:
a b c
0 2.0 3.0 1.0
1 1.0 3.0 2.0
2 2.0 1.0 3.0
3 2.0 3.0 1.0
带有重复值的轴索引值的Series
In [167]: obj = Series(range(5), index=['a', 'a', 'b', 'b', 'c'])
In [168]: obj
Out[168]:
a 0
a 1
b 2
b 3
c 4
dtype: int32
is_unique属性可以告诉它的值是否是唯一的
In [169]: obj.index.is_unique
Out[169]: False
对于带有重复值的索引,数据选取的行为将会有些不同。如果某个索引对应多个值,则返回一个Series;而对应单个值的,则返回一个标量值
In [170]: obj['a']
Out[170]:
a 0
a 1
dtype: int32
In [171]: obj['c']
Out[171]: 4
In [172]: df = DataFrame(np.random.randn(4, 3), index=['a', 'a', 'b', 'b'])
In [173]: df
Out[173]:
0 1 2
a -0.199619 -0.871154 -0.674903
a 1.573516 0.558822 0.511055
b 0.029318 -0.654353 -0.682175
b -0.563794 1.756565 0.105016
In [174]: df.ix['b']
Out[174]:
0 1 2
b 0.029318 -0.654353 -0.682175
b -0.563794 1.756565 0.105016
通过练习认识到有些地方不能很好的理解,以后学习中慢慢理解各种函数的使用。