[SPARK-34060][SQL][FOLLOWUP] Preserve serializability of canonicalized CatalogTable

[MaxGekk]

What changes were proposed in this pull request?

Replace toMap by map(identity).toMap while getting canonicalized representation of CatalogTable. CatalogTable became not serializable after #31112 due to usage of filterKeys. The workaround was taken from scala/bug#7005.

Why are the changes needed?

This prevents the errors like:

[info]   org.apache.spark.SparkException: Job aborted due to stage failure: Task not serializable: java.io.NotSerializableException: scala.collection.immutable.MapLike$$anon$1
[info]   Cause: java.io.NotSerializableException: scala.collection.immutable.MapLike$$anon$1

Does this PR introduce any user-facing change?

Should not.

How was this patch tested?

By running the test suite affected by #31112:

$ build/sbt -Phive-2.3 -Phive-thriftserver "test:testOnly *AlterTableDropPartitionSuite"

[SparkQA]

Test build #134107 has started for PR 31197 at commit 81d6786.

[SparkQA]

Kubernetes integration test starting
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/38691/

[SparkQA]

Kubernetes integration test status failure
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/38691/

[dongjoon-hyun]

+1, LGTM.
According to the original patch, do we need this for master/3.1/3.0, @MaxGekk ?

[SparkQA]

Kubernetes integration test starting
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/38707/

[MaxGekk]

According to the original patch, do we need this for master/3.1/3.0 ...

@dongjoon-hyun Yes

[SparkQA]

Kubernetes integration test status failure
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/38707/

[SparkQA]

Test build #134124 has finished for PR 31197 at commit d26ea30.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[dongjoon-hyun]

Thank you, @MaxGekk . Merged to master/3.1/3.0.

[dongjoon-hyun]

Do you think if we can have a way for us to prevent this kind of regression, @cloud-fan , @HyukjinKwon , @MaxGekk ?

[dongjoon-hyun]

Also, cc @gatorsmile

[HyukjinKwon]

I think it's a bug fixed in Scala 2.13. Once we switch it to 2.13 by default and drop 2.12, it will prevent permanently I guess.

[dongjoon-hyun]

I think we cannot drop 2.12 easily in Apache Spark 3.x line like Hadoop 2.7.x.