CODEWITHSUNDEEP
pythonpandas
Funkeh-Monkeh
Funkeh-Monkeh

Reputation: 661

Pandas: Fill rows with values based on condition in other rows

I am trying to work out how to fill some rows with text based on other columns but am struggling to figure it out. Below is an example of a subset of data I have in a pandas data frame.

              Group       Name_X       Name_Y  Code Group_Code Group_Correct
0           Group_A   Company_PJ   Company_PJ  M001   1-00-002       Group_A
1       Group_A_Ltd   Company_PJ          NaN   NaN        NaN           NaN
2           Group_B  Company_XYZ          NaN   NaN        NaN           NaN
3         Group_GTY    Company_R    Company_R  M020   1-00-033     Group_GTY
4           Group_A   Company_PJ          NaN   NaN        NaN           NaN
5         Group_BGG   Company_VV   Company_VV  M023   1-00-233     Group_BGG
6           Group_B  Company_XYZ          NaN   NaN        NaN           NaN
7           Group_B  Company_XYZ  Company_XYZ  M003   1-00-003       Group_B
8   Group_B_Limited  Company_XYZ          NaN   NaN        NaN           NaN
9           Group_B  Company_DEF  Company_DEF  M004   1-00-006       Group_B
10         Group_B+  Company_DEF          NaN   NaN        NaN           NaN

And what I would like is to get something like the below. The text shown in red are the text that needs to be filled or changed based on conditions of text in other rows.

enter image description here

What's the easiest way I could go about doing it in python.

Upvotes: 1

Views: 1995

Answers (1)

BENY
BENY

Reputation: 323396

Using transform first

df=df.groupby('Name_x').transform('first')

If you have some column different for each row

l=['name1','name2'] # column different from each row
pd.concat([df.drop(l,1).groupby('Name_x').transform('first'),df[l]],axis=1).reindex(columns=df.columns)

Update : 

pd.concat([df.drop(l,1).groupby('Name_x').transform('first'),df[l.append('Name_x')]],axis=1).reindex(columns=df.columns)

Update2 : 

pd.concat([df.drop(l,1).groupby('Name_x').transform('first'),df[l],df[['Name_x']]],axis=1).reindex(columns=df.columns)

Upvotes: 2

Related Questions