Multi tool use

Clash Royale CLAN TAG

How to keep pandas group by column when applying transform function?

This is my pandas dataframe look's like:

sampling_time MQ2_LPG MQ2_CO MQ2_SMOKE MQ2_ALCOHOL MQ2_CH4 MQ2_H2 MQ2_PROPANE
0 2018-07-15 08:41:49.028 4.41 32.87 19.12 7.70 10.29 7.59 4.49
1 2018-07-15 08:41:49.028 2.98 19.08 12.47 4.72 6.34 5.15 3.02
2 2018-07-15 08:41:49.028 2.73 16.88 11.33 4.22 5.69 4.72 2.76
3 2018-07-15 08:41:49.028 2.69 16.47 11.11 4.13 5.57 4.64 2.71
4 2018-07-15 08:41:49.028 2.66 16.26 11.00 4.09 5.50 4.60 2.69


When I'm doing group by (split apply combine method), my sampling time column was removed.

transformed = dataframe.groupby('sampling_time').transform(lambda x: (x - x.mean()) / x.std())
transformed.head()

MQ2_LPG MQ2_CO MQ2_SMOKE MQ2_ALCOHOL MQ2_CH4 MQ2_H2 MQ2_PROPANE
0 15.710127 15.975636 15.773724 15.876433 15.874190 15.694674
1 3.519619 3.313661 3.494836 3.408578 3.404160 3.563717
2 1.388411 1.293621 1.389884 1.316656 1.352130 1.425885
3 1.047418 0.917159 0.983665 0.940110 0.973294 1.028148
4 0.791673 0.724337 0.780556 0.772756 0.752306 0.829280


Any help or suggestion about how to keep the sampling time column would be very appreciated.

1 Answer
1

You can do this by setting 'sampling_time' into the index, then when you runs groupby with transform, you will get your transform columns out with the index.

df1 = df.set_index('sampling_time')

df1.groupby('sampling_time').transform(lambda x: x-x.std())


output:

MQ2_LPG MQ2_CO MQ2_SMOKE MQ2_ALCOHOL
sampling_time
2018-07-15 08:41:49.028 3.663522 25.760508 15.652432 6.154209
2018-07-15 08:41:49.028 2.233522 11.970508 9.002432 3.174209
2018-07-15 08:41:49.028 1.983522 9.770508 7.862432 2.674209
2018-07-15 08:41:49.028 1.943522 9.360508 7.642432 2.584209
2018-07-15 08:41:49.028 1.913522 9.150508 7.532432 2.544209

MQ2_CH4 MQ2_H2 MQ2_PROPANE
sampling_time
2018-07-15 08:41:49.028 8.243523 6.313227 3.7205
2018-07-15 08:41:49.028 4.293523 3.873227 2.2505
2018-07-15 08:41:49.028 3.643523 3.443227 1.9905
2018-07-15 08:41:49.028 3.523523 3.363227 1.9405
2018-07-15 08:41:49.028 3.453523 3.323227 1.9205


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.