CODEWITHSUNDEEP
pythonpandas
Tatik
Tatik

Reputation: 1227

How to check a type of column values in pandas DataFrame

I can check column types using df.dtypes, where df is pandas DataFrame. However, my question is a bit different. I have the following DataFrame:

col1 col2
0    <class 'pandas._libs.tslibs.timestamps.Timestamp'>
1    <class 'pandas._libs.tslibs.timestamps.Timestamp'>
2    <class 'float'>
3    NaN
4    <class 'pandas._libs.tslibs.timestamps.Timestamp'>

The df["col2"].dtypes returns object.

I need to create a new column is_timestamp that would check if col2 value is pandas timestamp. For testing, I tried this code:

isinstance(df_viz["col2"][0], pd._libs.tslibs.timestamps.Timestamp)

But it returns False.

The expected output:

col1 col2                                                 col3
0    <class 'pandas._libs.tslibs.timestamps.Timestamp'>   Yes
1    <class 'pandas._libs.tslibs.timestamps.Timestamp'>   Yes
2    <class 'float'>                                      No
3    NaN                                                  No
4    <class 'pandas._libs.tslibs.timestamps.Timestamp'>   Yes

Upvotes: 2

Views: 3199

Answers (2)

Thotsaphon Sirikutta
Thotsaphon Sirikutta

Reputation: 128

you can check for each row like this

df['check_datetime'] = [type(val) == datetime.datetime for val in df['datime_field']]

I'm not sure about your type you can find your type by (type(val)) and place them into code

if you want 'YES' and 'NO'

can using 

df['my_col'] = np.where(df['my_col'] is True,"YES","NO)

my try code

Upvotes: 1

anky
anky

Reputation: 75080

Try with:

df_viz['col3']=(df_viz.col2.transform(lambda x:
  np.where(x==pd._libs.tslibs.timestamps.Timestamp,'Yes','No')))

Upvotes: 2

Related Questions