CODEWITHSUNDEEP
pythonpandas
Bjc51192
Bjc51192

Reputation: 327

apply.pd.to_date_Time. apply being unmapped/ dealing with NaT

I'm a bit confused on an issue I have been having with my code test data below, as well as explanation. 

test = {"Med to Ind Date": ['', '', 1402531200000000000, '', 1402876800000000000], 
        "Med to Ind Indicator": ['', '', 'Y', '', 'Y']}
test = pd.DataFrame(test)
date_fields = ["Med to Ind Date"]
test.loc[:, date_fields] = test.loc[:, date_fields].apply(pd.to_datetime)

So when you run the above code you will see that all the blank time fields are mapped to NaT. Which is fine, however it is interrupting my code below:

if "Med to Ind Indicator" in test.columns:
    test["Med to Ind Indicator"] = np.where(test["Med to Ind Date"] != '', "Yes", '')

The code above looks at the Med to Ind Date field, and if it isn't blank it maps the column Med to Ind Indicator to Yes. The work around I had was trying to replace pd.NaT with "" which worked, but it in turn undid my date_time transformation, and returned it to the original form. Can you guys recommend an alternative? Also, how exactly does pandas view the NaT field?

Upvotes: 1

Views: 43

Answers (1)

Alexander
Alexander

Reputation: 109576

Use isnull() (or notnull()) to test for NaT:

np.where(test["Med to Ind Date"].isnull(), '', "Yes")

Resulting output of test:

       Med to Ind Date Med to Ind Indicator
0                 None                     
1                 None                     
2  1402531200000000000                  Yes
3                 None                     
4  1402876800000000000                  Yes

Upvotes: 2

Related Questions