Question

如何将单列切割为二维pandas DataFrame？

我有多个仪器的单列时间序列数据：

DATE        ISIN            ASK
01.05.2017  XS1431271250    0.2401
02.03.2017  XS1431271250    0.2578
01.03.2017  XS1431271250    0.2665
01.05.2017  XS1431274601    0.0474
02.03.2017  XS1431274601    0.0471
01.03.2017  XS1431274601    0.0479
01.05.2017  XS1431271094    0.2419
02.03.2017  XS1431271094    0.2454
01.03.2017  XS1431271094    0.254

而不是它我需要每个ISIN的单列：

Date        XS1431271250    XS1431274601    XS1431271094
01.05.2017  0.2401          0.0474          0.2419
02.03.2017  0.2578          0.0471          0.2454
01.03.2017  0.2665          0.0479          0.254

我尝试过分组，df.factorize，df.melt，但没有成功。请以正确的方式提出建议？

Answer 1

result = df.pivot('DATE','ISIN')
#                    ASK                          
#ISIN       XS1431271094 XS1431271250 XS1431274601
#DATE                                             
#01.03.2017       0.2540       0.2665       0.0479
#01.05.2017       0.2419       0.2401       0.0474
#02.03.2017       0.2454       0.2578       0.0471

如果您更喜欢扁平列名称，则可以删除＆＃39; ASK＆＃39;电平：

result.columns = result.columns.get_level_values(1)
#ISIN        XS1431271094  XS1431271250  XS1431274601
#DATE                                                
#01.03.2017        0.2540        0.2665        0.0479
#01.05.2017        0.2419        0.2401        0.0474
#02.03.2017        0.2454        0.2578        0.0471

Answer 2

此操作称为pivot：

>>> df.pivot(index='DATE', columns='ISIN')
                ASK                          
ISIN       XS1431271094 XS1431271250 XS1431274601
DATE                                             
01.03.2017       0.2540       0.2665       0.0479
01.05.2017       0.2419       0.2401       0.0474
02.03.2017       0.2454       0.2578       0.0471

https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.pivot.html

将单列切片到二维pandas DataFrame

2 个答案: