I'm trying to run my AWS Lambda function using the Python 3.13 runtime, but I'm hitting a wall. Whenever I attempt to import psycopg2, I get the following error: `Runtime.ImportModuleError: Unable to import module 'lambda_function': No module named 'psycopg2._psycopg'`. I've made a layer by downloading the binary version of psycopg2 specifically for this runtime, but no luck so far. I've searched through various resources but haven't found a solution. Does anyone have any advice on how to resolve this issue? Note: I can't downgrade the runtime.
4 Answers
I prefer not to use layers myself. Before deployment, I typically install the modules locally. You can use a command like this:
```bash
pip install --platform manylinux_2_17_aarch64 --python-version 3.13 --abi cp313 --only-binary=:all: -r requirements.txt -t .
```
Just ensure you're working in a separate directory for deployment purposes. This method has worked well for me.
Have you looked at the instructions provided by AWS? They have a specific guide on how to properly set up psycopg2 for Lambda with Python 3.13. Follow their steps; I managed to get it working without issues using that approach.
It sounds like the issue might be due to the structure of your Lambda layer. Make sure the necessary files are placed in the correct folder hierarchy inside the layer. Try checking the AWS documentation on packaging Python libraries for Lambda to see if that helps you sort it out.
While you mentioned browsing Stack Overflow and Reddit, did you try checking out the official AWS documentation? They have detailed instructions on importing the psycopg2 library for Lambda.
Thanks for the link! I'm going to try that out and see if it helps.