print elements in TimedeltaIndex

Question

In pandas I have a TimedeltaIndex named durations

print(durations)

TimedeltaIndex(['05:00:00', '11:15:00', '07:30:00'], dtype='timedelta64[ns]', name='Date', freq=None)

Is it possible just to print each element of the TimedeltaIndex individually?

For example if I print durations[0] this will output: Timedelta('0 days 05:00:00')

What would I need to print just 05:00:00 without any associated Timedelta information?

EDIT possible solution? Create empty list..

stuff = []
stuff = durations.index.astype(str).str[-8:-3]

Now it seems like I can print each element of the list:

stuff[0]
stuff[1]

returns:

'00:00'
'05:00'

So it's a little ambiguous. `timedelta` can store any unit of time; that display is only valid when you have times between `[0, 24Hours)`. Once you add a large timedelta like '25:01:15'` the display now shows as `'0 days 05:00:00'` for all of the timedeltas. The same issue arises with negative timedeltas... It's really difficult to know how to handle these cases. — ALollz, May 06 '20 at 15:59
There's are numerous proposed solutions: https://stackoverflow.com/questions/538666/format-timedelta-to-string. Many of the highly voted ones will fail with large or negative timedeltas so read carefully — ALollz, May 06 '20 at 16:10
I made an edit to the post, does this seem like a wonky way of creating a solution? — bbartling, May 06 '20 at 16:14

CHRD · Accepted Answer · 2020-05-06T16:27:16.743

1

You can use split to separate the different parts of the output and only choose the third item like below:

print(str(durations[0]).split(' ')[2])

edited May 06 '20 at 16:27

answered May 06 '20 at 15:51

CHRD

AttributeError: 'Timedelta' object has no attribute 'dt' – bbartling May 06 '20 at 15:54
Try just `durations[0].seconds.apply(lambda x: pd.Timedelta(seconds=x)) ` – CHRD May 06 '20 at 15:55
Thanks for the help, its still giving me an error. `AttributeError: 'int' object has no attribute 'apply'` – bbartling May 06 '20 at 15:58
dumb question can `durations[0]` be converted to a different data type like a float/string or something like that? – bbartling May 06 '20 at 16:00
I think the solution is here somewhere... https://stackoverflow.com/questions/51101432/timedelta-to-string-type-in-pandas-dataframe – bbartling May 06 '20 at 16:03
Happy to help :) – CHRD May 06 '20 at 16:16

1 Answers1