[SPARK-29172][SQL] Fix some exception issue of explain commands

[TomokoKomiyama]

What changes were proposed in this pull request?

Added try exception

Why are the changes needed?

The behaviors of run commands during exception handling are different depends on explain command. I think it should be unified.

[ >spark.sql("explain cost select * from hoge").show(false) ]

[ >spark.sql("explain extended select * from hoge").show(false) ]

Does this PR introduce any user-facing change?

No

How was this patch tested?

tested manually

[kiszk]

Indentation should be fixed

[TomokoKomiyama]

Thank you for the review.
I've done.

[srowen]

I'm not sure we want to let the explain plan proceed here? it has failed already, and I'm not sure about just appending it to the output as if successful.

[TomokoKomiyama]

@srowen
If we use 'extend', sql behave like as follows even if we have error

scala> spark.sql("explain extended select * from hoge").show(false)
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|plan                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|== Parsed Logical Plan ==
'Project [*]
+- 'UnresolvedRelation [hoge]

== Analyzed Logical Plan ==
org.apache.spark.sql.AnalysisException: Table or view not found: hoge; line 1 pos 31;
'Project [*]
+- 'UnresolvedRelation [hoge]

org.apache.spark.sql.AnalysisException: Table or view not found: hoge; line 1 pos 31;
'Project [*]
+- 'UnresolvedRelation [hoge]

== Optimized Logical Plan ==
org.apache.spark.sql.AnalysisException: Table or view not found: hoge; line 1 pos 31;
'Project [*]
+- 'UnresolvedRelation [hoge]

== Physical Plan ==
org.apache.spark.sql.AnalysisException: Table or view not found: hoge; line 1 pos 31;
'Project [*]
+- 'UnresolvedRelation [hoge]
|
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

, because of the try exception in sql/core/src/main/scala/org/apache/spark/sql/execution/QueryExecution.scala(L156-)

    } catch {
      case e: AnalysisException => e.toString
    }
    append(analyzedOutput)
    append("\n")
    QueryPlan.append(analyzed, append, verbose, addSuffix, maxFields)
    append("\n== Optimized Logical Plan ==\n")
    QueryPlan.append(optimizedPlan, append, verbose, addSuffix, maxFields)
    append("\n== Physical Plan ==\n")
    QueryPlan.append(executedPlan, append, verbose, addSuffix, maxFields)
  }

Other options such as cost, formated and codegen don't behave like this, so it needs something change to be consistent.
I suggest that add try exceptions in other explain commands to make these behaviors like 'extend' one.

[srowen]

OK, that's a good argument from consistency. @kiszk do you agree?

[kiszk]

I agree to make their behaviors in explain consistent.

[SparkQA]

Test build #4885 has finished for PR 25848 at commit 9430c12.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[srowen]

Merged to master