More idiomatic version of "df.isnull().any().any()" w/ a Pandas DataFrame?

Question

df.isnull().any().any()

This line evaluates to a boolean True/False, as it checks whether a Pandas dataframe contains any NaN's in its rows or columns. Is there a more concise/idiomatic way of checking this?

`not df.notnull().all().all()` ;) But seriously your version looks fine. — U2EF1, Jun 20 '14 at 03:12

score 4 · Accepted Answer · answered Jun 20 '14 at 06:03

4

I think it's to use numpy's any:

In [11]: df = pd.DataFrame([[1, 2], [3, np.nan]])

In [12]: df.isnull().any().any()
Out[12]: True

In [13]: np.any(df.isnull())
Out[13]: True

answered Jun 20 '14 at 06:03

Andy Hayden

328,850
93
598
514

This is much more intuitively readable :) thanks – maxm Jun 20 '14 at 12:27
1

@maxm ps. it's also faster/more efficient, as it's done in one pass with no temporaries. – Andy Hayden Jun 20 '14 at 18:30

More idiomatic version of "df.isnull().any().any()" w/ a Pandas DataFrame?

1 Answers1

Linked

Related