Open
Description
method df.merge will throw away / reset the index of df. I dont think this is desired result.
here's an example:
df = pd.DataFrame({'key' : [1,2,3]}, index=['very','important','index'])
df2 = pd.DataFrame({'key' : [3,2,1], 'data':['c','b','a']}, index=[100,200,300])
df.merge(df2, on='key')the result will lose the very important index in df and instead will have a generic 0 1 2 index.
key data
0 1 a
1 2 b
2 3 c
here is a method based on merge that will preseve the original index (which should be the intended result IMHO)
def merge_with_perfect_index(df, df2, on, how):
original_index_name = df.index.name
TEMP_INDEX = 'temp_index_aviad'
return df.rename_axis(TEMP_INDEX).reset_index().merge(df2, on=on, how=how).set_index(TEMP_INDEX).rename_axis(original_index_name)
merge_with_perfect_index(df, df2, on='key', how='left')which outputs
key data
very 1 a
important 2 b
index 3 c
tested on 1.5.3
Metadata
Metadata
Assignees
Labels
No labels