[python]pandas常用

import numpy as np
import pandas as pd

numpy -> pandas

# data_test_cur是一个pd.dataframe变量
df_pred = pd.DataFrame(data={'pop_pred':pop_pred})
df_pred = pd.DataFrame(data={'pop_pred':pop_pred},index=data_test_cur.index)

pandas -> numpy

a = df_pred.values

将两个df.dataframs给连在一起

#df_1的列是'column1','column2'
#df_2的列是'abc','bcd'
#结果是'column1','column2'，'abc','bcd'
data_test = df_1.join(df_2)

index

#season_dummy: 'mesh','abc','seaon_1','seaon_2','seaon_3','pop'
#a将取season_dummy的所有行，以及'seaon_1','seaon_2','seaon_3','pop'这几列
a = season_dummy.ix[:, 'season_1':]

#取'mesh', 'pop'这两列
seaon_dummy.columns = ['mesh', 'pop']
b = seaon_dummy[cols_to_keep]

#
seaon_dummy.columns[2:]

where

#season_dummy: 'mesh','abc','seaon_1','seaon_2','seaon_3','pop'
#取mesh为100的所有行
season_dummy[season_dummy['mesh']==100]