Shixiong Zhu

> [SPARK-41290](https://issues.apache.org/jira/browse/SPARK-41290) also contains changes required to #1072? No. It doesn't add the identity column syntax.

This is a complicated issue and we haven't figured out a solution. One workaround would be persisting the data after evaluating UDFs.

@sdaberdaku did you hit the same error? Do you have a quick reproduction we can try?

> insert into target_table (select from source_table) This is the issue I mentioned above `Currently Spark will block such INSERT query if the user doesn't provide the entire column list....

@keen85 [SPARK-41290](https://issues.apache.org/jira/browse/SPARK-41290) is the spark side ticket. We will make sure `insert into` works with Generated Columns when finishing the change in Spark side.

Could you enable faulthandler before running delta sharing code? ``` import faulthandler faulthandler.enable() ``` Then cancelling the code should trigger a thread dump and provide more useful information.

Could you share the thread dump for the fsspec call?

Closing this. We have published it to conda: https://anaconda.org/conda-forge/delta-spark

Hitting this issue as well. I think the root cause is [noExit](https://github.com/etsy/sbt-checkstyle-plugin/blob/58b5c5fc4fc2d45b56ea07a9ad858089e34853e7/src/main/scala/com/etsy/sbt/checkstyle/Checkstyle.scala#L129-L141) is not thread-safe. One thread can set the security manager back to the original one while another thread...

Sorry. I was not clear. I didn't ask to translate the various cases into specific exception types. I'm wondering if we can avoid wrapping the original cause with `FileNotFoundError`, such...