- 新しい順
- 投票が多い順
- コメントが多い順
Hi,
to successfully add an external library to a Glue Python Shell job you should follow the documentation at this link.
UPDATE as described i the link above, when using python 3.9 the best option to install external libraries is:
--additional-python-modules s3://aws-glue-native-spark/tests/j4.2/fbprophet-0.6-py3-none-any.whl,scikit-learn==0.21.3
For previous version the following is still correct considering you have already downloaded the wheel file and uploaded it to Amazon S3, then if you are creating your job via command line you need to add the parameter:
--default-arguments '{"--extra-py-files" : ["s3://MyBucket/python/library/openpyxl-3.0.9-py2.py3-none-any.whl"]}
if you are creating/editing the Python shell in the console:
-
for the new Glue Studio Job Editor : look under Job Details , Advanced properties.
-
for the legacy Job Editor - look under the Security configuration, script libraries, and job parameters (optional) section
Once you locate the text box under Python library path paste the full S3 URI for your wheel file.
I tested it with your library and it works in my environment.
Processing ./glue-python-libs-cr2dddvq/openpyxl-3.0.9-py2.py3-none-any.whl
Collecting et-xmlfile
Downloading et_xmlfile-1.1.0-py3-none-any.whl (4.7 kB)
Installing collected packages: et-xmlfile, openpyxl
Successfully installed et-xmlfile-1.1.0 openpyxl-3.0.9
hope this helps,
Thank you for your question. Without some context its hard to say what is the reason, but in general i was able to make it work as based on this article https://aws.amazon.com/premiumsupport/knowledge-center/glue-version2-external-python-libraries/
The article shared does not work either, right now I am importing just one library - openpyxl. Gives "No module named openpyxl" error. Have passed the wheel file downloaded from internet and added, also tried passing the key value job parameter (key:--additional-python-modules, value: openpyxl==3.0.9)
Hi, the answer is actually incorrect, the link provided works for AWS Glue Spark JObs , not for Glue Pyhon Shell as requested in the question.
it also could be improved by mentioning that it is possible to understand if an error is happening during the import of the external library by checking in the Cloudwatch logs for the job.
If no error are presents the logs under the job run will show the package installed.
関連するコンテンツ
AWS公式更新しました 2年前- AWS公式更新しました 2年前
I did the same thing but my problem was not solved. I have edited the Python shell in the console. In new version user interface of Jobs, not legacy version jobs, I found the Python library path in the library section in the Job details tab.