How can I quickly drop rows where A, B and C are all false? I tried:
df3 = df[~(~(df['A'])& ~(df['B']) & ~(df['C']) )]
df3
com A B C
0 as TRUE FALSE FALSE
1 da TRUE FALSE FALSE
Add a comment
|
3 Answers
drop rows where A, B and C are are all false
With df.sum across axis=1 along with comparison if sum in the row for these coumns is not equal to 0 , using df.ne
out = df[df[['A','B','C']].sum(1).ne(0)].copy()
4 Comments
asd
Thanks! Would you mind explaining this sorry? I didn't quite follow.
anky
@asd when you take sum of each row for columns
A,B,C , if there is any True value, it will return 1 or more , hence I check if the sum is not equal to zero (which means all columns for a row is False)asd
Thanks. Out of curiosity, what's wrong with the code I wrote? Am I misunderstanding some syntax or?
anky
@asd what you do is
False&False&False (which is False) ( assuming all the columns in the row is False) , then you invert it (~) which will always return True so for rows having all False values it still returns True and hence doesnot drop itWait a mins any
df[df[['A','B','C']].any(1)]
2 Comments
anky
This is the easiest built-in way..!!
Chris Adams
100% agree, I completely over-thought this one 😂
An alternative using all along axis 1. The 2nd ~ operator directly - directly before df, reversed all False to True. all returns True if all column values are True, then using ~ again to reverse this to index rows that are actually all False:
df3 = df[~(~df[['A', 'B', 'C']]).all(1)]