search

awslabs / python-deequ

Python API for Deequ
Apache License 2.0
731 stars 136 forks source link

Running pydeeque on EMR in cluster mode #119

Open prashant6700 opened 1 year ago

prashant6700 commented 1 year ago

Describe the bug We are trying to run pydeeque on an already running EMR cluster using ' spark-submit --deploy-mode cluster --py-files s3://path/pydeeque.zip --jars s3://path/deequ-2.0.1-spark-3.2.jar python_file.py ' However we land into following exception related to pandas

""" Traceback (most recent call last): File "python_file.py", line 5, in from pydeequ.verification import * File "/mnt/yarn/usercache/hadoop/appcache/application_1674190090846_0001/container_1674190090846_0001_02_000001/pydeequ.zip/pydeequ/init.py", line 19, in File "/mnt/yarn/usercache/hadoop/appcache/application_1674190090846_0001/container_1674190090846_0001_02_000001/pydeequ.zip/pydeequ/analyzers.py", line 9, in File "/mnt/yarn/usercache/hadoop/appcache/application_1674190090846_0001/container_1674190090846_0001_02_000001/pydeequ.zip/pydeequ/pandas_utils.py", line 8, in ModuleNotFoundError: No module named 'pandas' """

We also tried to zip and supply pandas with ' spark-submit --deploy-mode cluster --py-files s3://path/pydeeque.zip,s3://path/pandas.zip --jars s3://path/deequ-2.0.1-spark-3.2.jar python_file.py ' but landed into following issues

'''' Traceback (most recent call last): File "python_file.py", line 5, in from pydeequ.verification import * File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pydeequ/init.py", line 19, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pydeequ/analyzers.py", line 9, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pydeequ/pandas_utils.py", line 8, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pandas/init.py", line 22, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pandas/compat/init.py", line 15, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pandas/compat/numpy/init.py", line 7, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pandas/util/init.py", line 1, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pandas/util/_decorators.py", line 14, in File "/mnt1/yarn/usercache/hadoop/appcache/application_1674190090846_0002/container_1674190090846_0002_01_000001/pandas.zip/pandas/_libs/init.py", line 13, in ModuleNotFoundError: No module named 'pandas._libs.interval'

'''

To Reproduce Screenshot from 2023-01-25 17-02-01

EMR configurations: -Master: m5.xlarge, 1 instance -core:m5.xlarge, 1 instance

chenliu0831 commented 1 year ago

@prashant6700 I would suggest you try to get a simpler program who import Pandas working on EMR. It seems to be an environment setup issue not very relevant to PyDeequ