How to check if a Pandas row contains an empty set

Question

I want to check if a Pandas Dataframe row contains an empty set in a specific column, i.e.

d = {'col1': [1, 2], 'col2': [3, {}]}
df2 = pd.DataFrame(data=d)


    col1    col2
0   1       3
1   2       {}

and then

df2['col_2_contains_empty_set'] = ? #  how to implement this

should give

    col1    col2    col_2_contains_empty_set
0   1       3       False
1   2       {}      True

What's the correct way to do this? Can't do

bool(df['col2'])

or

df['col2'].bool()

as Series are have ambiguous Boolean values, I think.

@Sebastian Thanks for the point; I checked and https://stackoverflow.com/questions/36674083/why-is-it-possible-to-replace-set-with/36674133 seems to confirm that, in fact, sets *can* be declared like that. — Zubo, Dec 11 '17 at 23:23
Sets, yes, but not empty sets. Run `print(type({}))`, compared to `print(type({1}))`. — Sebastian Mendez, Dec 11 '17 at 23:31
@Sebastian Well, that's that then :) thanks for pointing that out! — Zubo, Dec 11 '17 at 23:33

Scott Boston · Answer 1 · 2017-12-11T23:04:22.740

6

One way:

df2.apply(lambda x: any(x.values == {}), axis=1)

Output:

0    False
1     True
dtype: bool

OR

df2['c'] = np.max(df2.values == {}, 1).astype(bool)

Output:

   col1 col2      c
0     1    3  False
1     2   {}   True

edited Dec 11 '17 at 23:04

answered Dec 11 '17 at 22:57

Scott Boston

147,308
15
139
187

score 3 · Answer 2 · answered May 10 '18 at 03:59

3

You can take advantage of the fact that len({})=0 and apply a lambda function:

df2['col2'].apply(lambda x: len(x)==0)

Note that this will return True for empty lists and dicts as well.

answered May 10 '18 at 03:59

John Santa Maria

31
1

score 2 · Answer 3 · answered Dec 11 '17 at 23:05

You can just compare df2.values to an empty dictionary:

In [ ]: df2['col_2_contains_empty_set'] = (df2.values == {}).any(axis=1)
   ...: df2
Out[ ]: 
   col1 col2  col_2_contains_empty_set
0     1    3                     False
1     2   {}                      True

BENY · Answer 4 · 2017-12-12T20:17:42.603

0

df2.applymap(type)==type({})
Out[1044]: 
    col1   col2
0  False  False
1  False   True

after assgin it back

df2['C']=(df2.applymap(type)==type({})).any(1)
df2
Out[1052]: 
   col1 col2      C
0     1    3   False
1     2   {}    True

edited Dec 12 '17 at 20:17

answered Dec 11 '17 at 23:53

BENY

317,841
20
164
234

How to check if a Pandas row contains an empty set

4 Answers4