[ERROR] No module named 'pyarrow._orc' in lambda python3.12

[debimishra89]

Describe the bug

I am using lambda python 3.12
awssdkpandas layer for python 3.12, version 16
I am trying to read an ORC file and create a dataframe. So i am using s3.read_orc API of wrangler package.
please advise.

How to Reproduce

1- create lambda function with python 3.12
2- add lambda layer by selecting AWSSDKPandas-3.12 ,version 16 from AWS layers.
3-import awswrangler
4- try to read orc file using s3.read_orc API.

Expected behavior

No response

Your project

No response

Screenshots

No response

OS

linux

Python version

3.12

AWS SDK for pandas version

AWSSDKPandas 3.12, version 16

Additional context

No response

[jaidisido]

ORC is not currently supported in our managed Lambda layers:
https://github.com/aws/aws-sdk-pandas/blob/main/building/lambda/build-lambda-layer.sh#L39

The reason some of the parameters in the Arrow build are turned off is to keep the size of the layer manageable and below the Lambda limit (250Mb unzipped, 50Mb zipped). ORC is one of them as it's less used compared to other formats (e.g. parquet, csv...).

[debimishra89]

Now I tried to read parquet files. it gives me below error