You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
#!/databricks/conda/envs/cudf-udf/bin/python3.1
# -*- coding: utf-8 -*-
import re
import sys
from pip._internal.cli.main import main
if __name__ == '__main__':
sys.argv[0] = re.sub(r'(-script\.pyw?|\.exe)?$', '', sys.argv[0])
sys.exit(main())
When installing manba via conda [conda install -c conda-forge mamba](https://github.com/NVIDIA/spark-rapids/blob/branch-22.04/jenkins/databricks/init_cudf_udf.sh#L30) , the default python=3.10 alone with pip-python=3.10 are installed, logs as below:
As our integration test runs with python=3.8. Then we tried to down-grade to [mamba install ... python=3.8 ](https://github.com/NVIDIA/spark-rapids/blob/branch-22.04/jenkins/databricks/init_cudf_udf.sh#L30), the mismatch occoured, logs as below:
$ mamba install -y python=3.8
The following packages will be downloaded:
- python 3.10.2 h85951f9_3_cpython installed
+ python 3.8.12 ha38a3c6_3_cpython conda-forge/linux-64 28MB
- python_abi 3.10 2_cp310 installed
+ python_abi 3.8 2_cp38 conda-forge/linux-64 4kB
There was no such issue before, as the default python was python =3.8 for conda/mamba package.
Describe the bug
Nightly build failed on Databricks with "pip: No such file or directory", due to the update of conda/mamba default to Python 3.10 from Python3.8
Steps/Code to reproduce bug
Run nightly build with init script "init_cudf_udf.sh"
The text was updated successfully, but these errors were encountered: