Selecting rows where column is not null and equal to a (string) value

Question

So basically I want select all rows where Col A is equal to the string 'hey'. My problem is that Col A can contain null/nan's therefore I get a

TypeError: invalid type comparison.

When executing:

df.loc[df['A'] == 'hey']

I then made another condition:

df.loc[df['A'].notnull() & (df['A'] == 'hey')]

Here i get the same error.

I made a hack where I change all the null values in Col A to '' but thats not beautiful is there anyway nice to first choose all the rows where Col A isn't null and then from there all the ones who are equal to 'hey'?

df.loc[df['A'] .astype(str) == 'hey'] is working. Thanks Jezrael — Thor Bjorn, Jun 27 '18 at 09:34

score 1 · Answer 1 · answered Jun 27 '18 at 09:27

1

I guess there should be some numeric, so try convert values to strings or compare numpy array:

newDf = df[df.A.astype(str) == 'hey']

Or:

newDf = df[df.A.values == 'hey']

answered Jun 27 '18 at 09:27

jezrael

822,522
95
1,334
1,252

Hey man! Thanks the conversion to .astype(str) worked for me! – Thor Bjorn Jun 27 '18 at 09:32

score 0 · Answer 2 · answered Jun 27 '18 at 09:26

0

How about this?

df['A'] = df['A'].astype(str)
newDf = df[df.A == 'hey']

This should give you a new dataframe with all rows that contained "hey" from column A?

answered Jun 27 '18 at 09:26

Ankur Sinha

6,473
7
42
73

No because you can't compare a string and a null value. Thats why my first solution wouldn't work. – Thor Bjorn Jun 27 '18 at 09:29

score 0 · Answer 3 · answered Jun 27 '18 at 09:51

For null / NaN values, your logic is fine. Below is an example. You should provide a minimal and verifiable example, indicating version numbers for Python / Pandas.

df = pd.DataFrame({'col': [np.nan, None, 'hey', 45.4352, 'somestring']})

print(df.loc[df['col'] == 'hey'])

   col
2  hey

Selecting rows where column is not null and equal to a (string) value

3 Answers3