Select rows from a DataFrame based on True or False in a column in pandas

Question

Select rows from a DataFrame based on True or False in a column in pandas:

For example,

import pandas as pd df = {'uid':["1", "1", "1", "1", "2", "2", "2", "2"], 'type': ["a", "a", "b", "a", "a", "b", "b", "a"], 'is_topup':["FALSE", "FALSE", "TRUE", "FALSE","FALSE", "TRUE", "TRUE", "FALSE"], 'label':["FALSE", "FALSE", "TRUE", "FALSE","FALSE", "TRUE", "TRUE", "FALSE"]} df = pd.DataFrame(df) uid type is_topup label 0 1 a FALSE FALSE 1 1 a FALSE FALSE 2 1 b TRUE TRUE 3 1 a FALSE FALSE 4 2 a FALSE FALSE 5 2 b TRUE TRUE 6 2 b TRUE TRUE 7 3 a FALSE FALSE 8 3 b TRUE TRUE 9 3 b TRUE TRUE 10 3 a FALSE FALSE

I want to select a row in conditions like is

 uid type is_topup label 0 1 a FALSE FALSE 1 1 a FALSE FALSE 2 1 b TRUE TRUE 4 2 a FALSE FALSE 5 2 b TRUE TRUE 7 3 a FALSE FALSE 8 3 b TRUE TRUE

I tried to look at pandas documentation but did not find the answer.

Can you explain which rows you want to drop. There doesn't seem to be a logical reason to drop them without you explaining which rows you want to drop. Why do I have to drop rows 3, 6, 9 and 10? Doesn't seem like droping duplicates (0 and 1 are duplicates). — Adirio
– Adirio, Commented Aug 27, 2019 at 7:33

Chris · Accepted Answer · 2019-08-27 07:38:33Z

Not sure the most efficient way but using idxmax:

new_df = df.groupby('uid').apply(lambda x: x[:(x['is_topup'] & x['label']).reset_index(drop=True).idxmax()+1]) print(new_df)

Output:

 uid type is_topup label uid 1 0 1 a False False 1 1 a False False 2 1 b True True 2 4 2 a False False 5 2 b True True 3 7 3 a False False 8 3 b True True

yes, but "unsupported operand type(s) for &: 'str' and 'bool'"
is is_topup and label strs? In such a case, replace (x['is_topup'] & x['label']) with (x['is_topup'].eq('TRUE') & x['label'].eq('TRUE'))

RaJa · Accepted Answer · 2019-08-27 07:20:33Z

0

It seems to me that a simple

result = df.drop_duplicates()

should do the trick. At least your given example would work that way.

answered Aug 27, 2019 at 7:20

RaJa

1,5971 gold badge13 silver badges18 bronze badges

1 Comment

artpii Over a year ago

not like that, I want all uid and some row. example uid= 3 new, with 4 row and I want 2 row(7,8) is_topup and label is change to TRUE TRUE

Collectives™ on Stack Overflow

Select rows from a DataFrame based on True or False in a column in pandas

2 Answers 2

2 Comments

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Related