"SELECT ((1) NOT BETWEEN (CASE ((\)) WHEN (1) THEN 0 END ) AND (\)) FROM \
" brings different results, when using CPU and GPU.
What you expected to happen:
It is the same result, when using CPU and GPU.
Minimal Complete Verifiable Example:
import pandas as pd
import dask.dataframe as dd
from dask_sql import Context
c = Context()
df0 = pd.DataFrame({
'c0': [True],
'c1': [3666.0000],
'c2': [0.3820597044277436],
})
t0 = dd.from_pandas(df0, npartitions=1)
c.create_table('t0', t0, gpu=False)
c.create_table('t0_gpu', t0, gpu=True)
print('CPU Result:')
result1= c.sql("SELECT ((1) NOT BETWEEN (CASE ((t0.c1)) WHEN (1) THEN 0 END ) AND (t0.c0)) FROM t0").compute()
print(result1)
print('GPU Result:')
result2= c.sql("SELECT ((1) NOT BETWEEN (CASE ((t0_gpu.c1)) WHEN (1) THEN 0 END ) AND (t0_gpu.c0)) FROM t0_gpu").compute()
print(result2)
Result:
INFO:numba.cuda.cudadrv.driver:init
CPU Result:
Int64(1) NOT BETWEEN CASE t0.c1 WHEN Int64(1) THEN Int64(0) END AND t0.c0
0 False
GPU Result:
Int64(1) NOT BETWEEN CASE t0_gpu.c1 WHEN Int64(1) THEN Int64(0) END AND t0_gpu.c0
0 <NA>
INFO:numba.cuda.cudadrv.driver:add pending dealloc: module_unload ? bytes
This problem came up at dask-sql version: 2023.6.0 .
And it has been fixed at dask-sql version: 2024.3.0, after my verification.
Thanks to the developers for their contributions.
What happened:
"SELECT ((1) NOT BETWEEN (CASE ((\)) WHEN (1) THEN 0 END ) AND (\)) FROM \" brings different results, when using CPU and GPU.
- dask-sql version: 2023.6.0
- Python version: Python 3.10.11
- Operating System: Ubuntu22.04
- Install method (conda, pip, source): Docker deploy by https://hub.docker.com/layers/rapidsai/rapidsai-dev/23.06-cuda11.8-devel-ubuntu22.04-py3.10/images/sha256-cfbb61fdf7227b090a435a2e758114f3f1c31872ed8dbd96e5e564bb5fd184a7?context=explore
qwebug
commented
5 months ago
- © Githubissues.
- Githubissues is a development platform for aggregating issues.
What you expected to happen:
It is the same result, when using CPU and GPU.
Minimal Complete Verifiable Example:
Result:
Anything else we need to know?:
Environment:
This problem came up at dask-sql version: 2023.6.0 . And it has been fixed at dask-sql version: 2024.3.0, after my verification. Thanks to the developers for their contributions.