标签：class IV creat 建议 日期 date 重复 title 取数
.loc.ilocIn [73]: s = pd.Series([1,3,5,np.nan,6,8])
In [74]: s
Out[74]:
0 1.0
1 3.0
2 5.0
3 NaN
4 6.0
5 8.0
dtype: float64
In [75]: dates = pd.date_range(‘20130101‘, periods=6)
In [76]: df = pd.DataFrame(np.random.randn(6,4), index=dates, columns=list(‘ABCD‘))
In [77]: df
Out[77]:
A B C D
2013-01-01 -0.411674 0.273549 0.629843 1.881497
2013-01-02 1.240512 0.970725 0.033099 1.553420
2013-01-03 -0.544326 0.545738 -1.325810 0.130738
2013-01-04 1.044803 -0.117151 0.874583 2.278227
2013-01-05 -2.194728 -2.536257 0.478644 0.057728
2013-01-06 -1.092031 1.249952 1.598761 -0.153423
#---pd.date_range?---
In [115]: pd.date_range(start=‘12/31/2011‘, end=‘12/31/2013‘, freq=‘A‘)
Out[115]: DatetimeIndex([‘2011-12-31‘, ‘2012-12-31‘, ‘2013-12-31‘], dtype=‘datetime64[ns]‘, freq=‘A-DEC‘)
In [78]: 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‘ })
...: df2
...:
Out[78]:
A B C D E F
0 1.0 2013-01-02 1.0 3 test foo
1 1.0 2013-01-02 1.0 3 train foo
2 1.0 2013-01-02 1.0 3 test foo
3 1.0 2013-01-02 1.0 3 train foo
In [80]: df2.dtypes
Out[80]:
A float64
B datetime64[ns]
C float32
D int32
E category
F object
dtype: object
在ipython中可以使用<tab>键进行自动补充，它会列出数据对象可以执行的操作。
df.head()
df.tail(3)
df.index
df.columns #返回一个这样的东西：pandas.indexes.numeric.Int64Index
df.values #提取出数据框的数值，返回一个array
建议 使用pandas的数据选取方法：.at, .iat, .loc, .iloc, .ix. 这些更高效。
df[‘A‘] # 选取某一列，返回一个Series，== df.A，【只能选某一列，不能用":"多选。】
df[0:3] # 选行
df[‘20130102‘:‘20130104‘]
通过标签label选取，.loc
用.loc[]选取数据时，方括号里对应的是：[行,列]（逗号分隔），如果只有一个值，默认是行。可以用“:”。
In [82]: df
Out[82]:
A B C D
2013-01-01 -0.411674 0.273549 0.629843 1.881497
2013-01-02 1.240512 0.970725 0.033099 1.553420
2013-01-03 -0.544326 0.545738 -1.325810 0.130738
2013-01-04 1.044803 -0.117151 0.874583 2.278227
2013-01-05 -2.194728 -2.536257 0.478644 0.057728
2013-01-06 -1.092031 1.249952 1.598761 -0.153423
In [83]: df.loc[dates[0]] # 作为index的日期列叫dates
Out[83]:
A -0.411674
B 0.273549
C 0.629843
D 1.881497
Name: 2013-01-01 00:00:00, dtype: float64
#---对多个维度轴axis进行选取---
In [84]: df.loc[‘20130102‘:‘20130104‘,[‘A‘,‘B‘]]
Out[84]:
A B
2013-01-02 1.240512 0.970725
2013-01-03 -0.544326 0.545738
2013-01-04 1.044803 -0.117151
#---选取某个数值---
In [85]: df.loc[dates[0],‘A‘]
Out[85]: -0.41167416696608039
In [86]: df.at[dates[0],‘A‘] # 更高效的做法
Out[86]: -0.41167416696608039
多重索引的选取
index有多个维度
#这里有一个多重索引
MultiIndex(levels=[[1, 2, 3], [‘count‘, ‘mean‘, ‘std‘, ‘min‘, ‘5%‘, ‘10%‘, ‘15.0%‘, ‘20%‘, ‘25%‘,
‘30.0%‘, ‘35%‘, ‘40%‘, ‘45%‘, ‘50%‘, ‘55.0%‘, ‘60.0%‘, ‘65%‘, ‘70%‘,
‘75%‘, ‘80%‘, ‘85.0%‘, ‘90%‘, ‘95%‘, ‘max‘]],
labels=[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2],
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
21, 22, 23, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
18, 19, 20, 21, 22, 23, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
15, 16, 17, 18, 19, 20, 21, 22, 23]],
names=[‘label_1‘, None])
df[columnName] #选某一列，或多列（":",[,,,]）
df.loc[:,columnName] #选某一列，或多列（":",[,,,]）
df.loc[1,columnName] #可以直接用最外层的索引
df.loc[(1,‘std‘),columnName] #多层索引要用tuple。选多行用":"连接tuple。
df.loc[[(1,‘std‘),(2,"count")],‘feature_001‘]
用位置选取：.iloc
.lic[]，位置索引，方括号里是整数值。同样的用“,”隔开行列。
In [93]: df.iloc[3]
Out[93]:
A 1.044803
B -0.117151
C 0.874583
D 2.278227
Name: 2013-01-04 00:00:00, dtype: float64
In [94]: df.iloc[3:5,0:2]
Out[94]:
A B
2013-01-04 1.044803 -0.117151
2013-01-05 -2.194728 -2.536257
In [95]: df.iat[1,1]
Out[95]: 0.97072539301549565
**布尔索引 **Boolean Indexing
某一列大于0的数据
In [96]: df[df.A > 0]
Out[96]:
A B C D
2013-01-02 1.240512 0.970725 0.033099 1.553420
2013-01-04 1.044803 -0.117151 0.874583 2.278227
整体大于零的数据。小于0的为NaN
In [97]: df[df > 0]
Out[97]:
A B C D
2013-01-01 NaN 0.273549 0.629843 1.881497
2013-01-02 1.240512 0.970725 0.033099 1.553420
2013-01-03 NaN 0.545738 NaN 0.130738
2013-01-04 1.044803 NaN 0.874583 2.278227
2013-01-05 NaN NaN 0.478644 0.057728
2013-01-06 NaN 1.249952 1.598761 NaN
对字符型数据选取
#---isin ---
In [98]: df2 = df.copy()
...: df2[‘E‘] = [‘one‘, ‘one‘,‘two‘,‘three‘,‘four‘,‘three‘]
...: df2
...:
Out[98]:
A B C D E
2013-01-01 -0.411674 0.273549 0.629843 1.881497 one
2013-01-02 1.240512 0.970725 0.033099 1.553420 one
2013-01-03 -0.544326 0.545738 -1.325810 0.130738 two
2013-01-04 1.044803 -0.117151 0.874583 2.278227 three
2013-01-05 -2.194728 -2.536257 0.478644 0.057728 four
2013-01-06 -1.092031 1.249952 1.598761 -0.153423 three
In [99]: df2[df2[‘E‘].isin([‘two‘,‘four‘])]
Out[99]:
A B C D E
2013-01-03 -0.544326 0.545738 -1.325810 0.130738 two
2013-01-05 -2.194728 -2.536257 0.478644 0.057728 four
使用布尔面具
In [107]: mask = df2["A"] >0
In [108]: df3 = df2[mask]
In [109]: df3
Out[109]:
A B C D E
2013-01-02 1.240512 0.970725 0.033099 1.553420 ONE
2013-01-04 1.044803 -0.117151 0.874583 2.278227 THREE
# 查看无重复的值：.unique()
In [101]: df2.loc[:,"E"].unique()
Out[101]: array([‘one‘, ‘two‘, ‘three‘, ‘four‘], dtype=object)标签：class IV creat 建议 日期 date 重复 title 取数
原文地址：https://www.cnblogs.com/pejsidney/p/9226613.html
