熊猫用另一个更新一个数据框

Question

我有2个数据框：

df1:

col1 col2 col3 

A     B    C


df2:
col1 col2 col3 col4
A     B    C    D

我想用df2更新df1，以便如果我在col1，col2，col3上匹配，那么用col4替换col3

df_want:
col1 col2 col3
A     B    D

在SQL中，它将是：

select df1.col1,df1.col2,coalesce( df2.col4,df1.col3) as col3
from df1 left join df2 on df1.col1=df2.col1 and df1.col2=df2.col2 and df1.col3=df2.col3

Answer 1

尝试：

df3 = df1.merge(df2, on=['col1', 'col2', 'col3'])
del df3['col3']
df3.columns = ['col1','col2','col3']

Answer 2

使用merge

df1=df1.merge(df2,how='left')
df1.col3=df1.col4.fillna(df1.col3)
df1
Out[189]: 
  col1 col2 col3 col4
0    A    B    D    D
df1.drop('col4',1,inplace=True)
df1
Out[191]: 
  col1 col2 col3
0    A    B    D

Answer 3

尝试一下：

for index, row in df1.iterrows():
    if row['col1']==df2.ix[index, 'col1'] and row['col2']==df2.ix[index, 'col2'] and row['col3']==df2.ix[index, 'col3']:
        row[C]=df2.ix[index, 'col4']