Python Cloud Function deploy successfully but doesn't work when triggered by scheduler/pub-sub

[Vkzera]

Hi,

I'm trying do schedule a cloud function, that worked perfectly when I tested and deployed, but it is giving me an error. The error happens in a specific part, when I'm trying to use pandas df.to_excel:

df = client.query(query).to_dataframe() -> Until here, works perfectly

xlsx_file = 'xlsx_file.xlsx'

df = df.to_excel(xlsx_file, index=False') -> Gives me the error:

OSError: [Errno 30] Read-only file system: 'excel_file.xlsx'

I've tried to change the engine from openpyxl to xlsxwritter, but also failed. What I don't get is that works when testing and deploying but doesn't with scheduler - pub/sub.

Thanks

[christianpaula]

Hi @Vkzera,

Welcome to Google Cloud Community!

Your Cloud Function is encountering a "Read-only file system" error because it lacks permission to write to the designated location. Here are the key steps to resolve it:

1. Choose a writable location:

• Cloud Storage: Create a Cloud Storage bucket and write the Excel file there using the google-cloud-storage library.
• Cloud Function's temporary storage: Write to the tmp directory, but remember these files are transient.
• Alternative formats: Consider using CSV or other supported formats if neither of the above work.

2. Ensure proper environment:

• Dependencies: Include necessary libraries like Pandas in the deployment package.
• Permissions: Verify the executing user has write permissions in the chosen directory.

3. Debug and refine:

• Use detailed logging and Stackdriver reporting to pinpoint the issue.
• Test locally to replicate the error under similar conditions.
• Consider file size limits and function timeout constraints.