Closed
Description
Describe the bug
While performing operations like shift on a SeriesGroupBy objects, the correct columns are not being picked up.
Steps/Code to reproduce bug
tmp.csv
>>> import cudf
>>> df = cudf.read_csv("tmp.csv")
>>> df
Unnamed: 0 Unnamed: 0.1 days_from_start hours_from_start categorical_id power_usage hour day_of_week _id_
0 120 5664 1332 2.001283 1 3.313606 1 1 0
1 409 5665 1332 2.001942 1 3.160683 2 1 0
2 1004 5666 1332 2.002601 1 3.160683 3 1 0
3 1356 5667 1332 2.003260 1 3.313606 4 1 0
4 1767 5668 1332 2.003920 1 3.237144 5 1 0
.. ... ... ... ... ... ... ... ... ...
331 122219 5995 1345 2.219437 1 3.237144 20 7 0
332 122699 5996 1345 2.220096 1 0.025749 21 7 0
333 122910 5997 1345 2.220755 1 1.325600 22 7 0
334 123417 5998 1345 2.221415 1 3.237144 23 7 0
335 123737 5999 1345 2.222074 1 3.237144 24 7 0
[336 rows x 9 columns]
>>> df.groupby("_id_")['power_usage'].shift(0)
/nvme/0/pgali/envs/cudfdev/lib/python3.8/site-packages/cudf/core/frame.py:2986: FutureWarning: keep_index is deprecated and will be removed in the future.
warnings.warn(
0 120
1 409
2 1004
3 1356
4 1767
...
331 122219
332 122699
333 122910
334 123417
335 123737
Name: Unnamed: 0, Length: 336, dtype: int64
>>> import pandas as pd
>>> pdf = pd.read_csv("tmp.csv")
>>> pdf.groupby("_id_")['power_usage'].shift(0)
0 3.313606
1 3.160683
2 3.160683
3 3.313606
4 3.237144
...
331 3.237144
332 0.025749
333 1.325600
334 3.237144
335 3.237144
Name: power_usage, Length: 336, dtype: float64Expected behavior
shift should happen on the correct column similar to pandas.
Environment overview (please complete the following information)
- Environment location: [Bare-metal, Docker, Cloud(specify cloud provider)]
- Method of cuDF install: [conda, Docker, or from source]
- If method of install is [Docker], provide
docker pull&docker runcommands used
- If method of install is [Docker], provide
Environment details
Please run and paste the output of the cudf/print_env.sh script here, to gather any other relevant environment details
Additional context
Add any other context about the problem here.