_sqldf not defined on Azure job cluster v12.2

[piterpan]

Since yesterday we have errors in notebooks that were previously working.

NameError: name '_sqldf' is not defined

It was working previously.

We are on Azure databricks, usng job pool 
Driver: Standard_D4s_v5 · Workers: Standard_D4s_v5 · 1-6 workers · from pool · 12.2 LTS

We have opened a ticket on Azure, no response for now 😕 

[Kale-S]

I have the same error since yesterday. 

I have rewrote all my magic SQL (%SQL) cells with the spark function:
query = """
              select * 
              from table
"""

df = spark.sql(query)

[samirsaliba]

Same here, also on Azure Databricks. Cluster version is 12.2 LTS Photon.

[ViniciusTavares]

Today I had the same error, the df _sqldf stopped working

[aholts]

Having same issue since Thursday

[whitebulb]

Same here. It was working completely fine a couple weeks ago.

[Tharun-Kumar · Accepted solution]

@piterpan 

This was a regression issue which impacted the jobs where _sqldf was referenced and the notebook those weren't run interactively. Our Engineering team has fixed this issue yesterday.

Could you check whether you are still facing the issue?

[piterpan]

Thank for your feedback.
We noticed it's working on v13.2 and we have been using v13.2 as a workaround.
I am waiting for the test on v12.2 and I'll confirm you if it's resolved for us.

[piterpan]

I can confirm it's working again.