marking same across rows if one of rows satisfy condition-CodePudding

How can I mark as 'abuser' across rows of same ID if one of rows of that ID satisfy a condition?

For example, if I have the following table,

ID	social score
1	0
1	2
1	3
2	3
2	1
2	2

Can I mark rows of ID 1 as all abnormal (because social score of one of rows is 0 ) and mark rows of ID 2 as all normal (because social score of none of rows of ID 2 is 0 )?

ID	social score	class
1	0	'abnormal'
1	2	'abnormal'
1	3	'abnormal'
2	3	'normal'
2	1	'normal'
2	2	'normal'

If I can, with what python phrase?

CodePudding user response：

First find the ID's having social score 0, then use np.where to assign class values:

i = df.loc[df['social score'].eq(0), 'ID']
df['class'] = np.where(df['ID'].isin(i), 'abnormal', 'normal')

   ID  social score     class
0   1             0  abnormal
1   1             2  abnormal
2   1             3  abnormal
3   2             3    normal
4   2             1    normal
5   2             2    normal

CodePudding user response：

Try this

df['class'] = df.ID.apply(lambda x: 'abnormal' if x==1 else 'normal')

CodePudding user response：

You can use groupby then use transform and at the end, use map and dict to replace True, False with abnormal, normal:

import pandas as pd

df = pd.DataFrame({'ID':[1,1,1,2,2,2],'social score': [0,2,3,3,1,2]})

df['class'] = df.groupby(['ID'])['social score'].transform(
    lambda x : any(x.eq(0))).map({True:'abnornmal',False:'normal'})
print(df)

Output:

   ID  social score      class
0   1             0  abnornmal
1   1             2  abnornmal
2   1             3  abnornmal
3   2             3     normal
4   2             1     normal
5   2             2     normal