Pyspark command can't be parsed

SonarQube Cloud
1

Hello everyone,

I’ve been having an issue with the analysis of a pyspark file:

  • Azure Devops used
  • Scanner used: sonar-scanner cli
  • Languages of the repository: python (since pyspark is treated as python)

this is the error that I have:

2026-10-30T09:38:17.9237841Z 10:38:13.883 INFO  19 source files to be analyzed
2026-10-30T09:38:17.9257597Z ##[error]10:38:16.508 ERROR Unable to parse file: Fabric/workspace/RawToBronze.Notebook/notebook-content.py
2026-10-30T09:38:17.9263838Z 10:38:16.508 ERROR Unable to parse file: Fabric/workspace/RawToBronze.Notebook/notebook-content.py
2026-10-30T09:38:17.9266393Z ##[error]10:38:16.510 ERROR Parse error at line 65 column 0:

   44: load_id = "2206100"
   45: keyvault_name = ""
   46: src_system = "SAP_ERP"
   47: src_name = ""
   48: bronze_load_mode = None
   49: 
   50: 
   51: 
   52: 
   53: 
   54: 
   55: 
   56: 
   57: 
   58: 
   59: 
   60: 
   61: 
   62: 
   63: 
   64: 
  -->  %run SDE_NB_000_Utils
   66: 
   67: 
   68: 
   69: 
   70: 
   71: 
   72: 
   73: 
   74: 
   75: 
   76: olu = OneLakeUtils()
   77: 
   78: 
   79: 
   80: 
   81: 
   82: 
   83: 
   84: 
   85: 
   86: 
   87: 
   88: 
   89: 
   90: 
   91: du = DeltaUtils(spark, keyvault_name, src_system, src_name)
   92: raw_data_root_dir = du.get_raw_data_root_dir()
2026-10-30T09:38:17.9267857Z 10:38:16.510 ERROR Parse error at line 65 column 0:
2

Hey @Giuseppe_Rizza, is this actually a pure python file? Or does it contain notebook syntax? I see the %run command, which is typical of notebooks but is not real python syntax. I suspect the process by which you are exporting the notebook into a .py file isn’t producing a 100% valid python file.

3

Yes, I confirm it’s a notebook.py file

4

Okay, I see. In principle, we should be able to analyze notebook files. However, I’m not sure we expect them to have a .py file extension. I’m going to flag this for the team to verify.

6

Hello @Giuseppe_Rizza.

right now we don’t yet support Notebook in normal .py files. We have a ticket to detect those usages. At this moment, the workaround would be to provide the notebook in Jupyter .ipynb format. Fabric supports exporting notebooks in that format.

Cheers,
Victor

7

thank you so much for the support!