Get previous month average for each day in Pandas

Question

I am trying to get average values for last 30 days.

That means that for row #3 it would be average of Close values between 2017-12-15 00:02:00 and 2018-01-13 00:03:00 1360.618915. So far I created helping columns called minusMonth and plusDay.

Result should be average of Close values between days minusMonth and plusDay. However right now it counts same value for each period

gData['result']=gData[(gData['Local time']>gData['minusMonth']) & (gData['Local time']<gData['plusDay'])]['Close'].mean()

Local time  Close   minusMonth  plusDay result
0   2018-01-12 00:00:00 1323.149    2017-12-15 00:00:00 2018-01-13 00:00:00 1360.618915
1   2018-01-12 00:01:00 1322.958    2017-12-15 00:01:00 2018-01-13 00:01:00 1360.618915
2   2018-01-12 00:02:00 1322.951    2017-12-15 00:02:00 2018-01-13 00:02:00 1360.618915
3   2018-01-12 00:03:00 1322.821    2017-12-15 00:03:00 2018-01-13 00:03:00 1360.618915
4   2018-01-12 00:04:00 1322.801    2017-12-15 00:04:00 2018-01-13 00:04:00 1360.618915
... ... ... ... ... ...
1172095 2020-04-04 23:55:00 1617.048    2020-03-07 23:55:00 2020-04-05 23:55:00 1360.618915
1172096 2020-04-04 23:56:00 1617.048    2020-03-07 23:56:00 2020-04-05 23:56:00 1360.618915
1172097 2020-04-04 23:57:00 1617.048    2020-03-07 23:57:00 2020-04-05 23:57:00 1360.618915
1172098 2020-04-04 23:58:00 1617.048    2020-03-07 23:58:00 2020-04-05 23:58:00 1360.618915
1172099 2020-04-04 23:59:00 1617.048    2020-03-07 23:59:00 2020-04-05 23:59:00 1360.618915
1172100 rows × 9 columns

When I do it with text instead of column reference, it works.

[IN] gData[(gData['Local time']>'2018.05.20') & (gData['Local time']<'2018.06.20')]['Close'].mean()
[OUT]1294.6491271981902

I have been stucked on this couple days, hopefully somebody can help.

Answer 1

You can use lambda function to apply on the entire dataframe with axis = 1 since it should be applied to every row of the dataframe.

df['result'] = df.apply(lambda x: df[(x['Local time']>df['minusMonth']) & (x['Local time']<df['plusDay'])]['Close'].mean(), axis = 1)