I'm trying to change the format of a datetime column in my Dataframe using lambda and strftime like below
df['Date Column'] = df['Date Column'].map(lambda x: x.strftime('%m/%d/%Y'))
However, since I have null values in some of these fields, this is giving me an error. I cannot drop these null rows because I still need them for the data in the other columns. Is there a way around this error without dropping the nulls.
Perhaps something like
df['Date Column'].map(lambda x: x.strftime('%m/%d/%Y') if x != null else "")
?
The method I've used is to drop the nulls, format the column, then merge it back onto the original dataset, but this seems like a very inefficient method.
You should be not checking for nan/nat (un)equality, but .notnull() should work and it does for me:
s = pd.date_range('2000-01-01', periods=5).to_series().reset_index(drop=True)
s[2] = None
s
0 2000-01-01
1 2000-01-02
2 NaT
3 2000-01-04
4 2000-01-05
dtype: datetime64[ns]
s.map(lambda x: x.strftime('%m/%d/%Y') if pd.notnull(x) else '')
0 01/01/2000
1 01/02/2000
2
3 01/04/2000
4 01/05/2000
dtype: object
This returns the same that the answers by @Alexander and @Batman but is more explicit. It may also be slightly slower for large series.
Alternatively you can use the .dt accesor. The null values will be formatted as NaT.
s.dt.strftime('%m/%d/%Y')
0 01/01/2000
1 01/02/2000
2 NaT
3 01/04/2000
4 01/05/2000
dtype: object