[SPARK-15907] [SQL] Issue Exceptions when Not Enough Input Columns for Dynamic Partitioning

[gatorsmile]

What changes were proposed in this pull request?

CREATE TABLE table_with_partition(c1 string)
PARTITIONED by (p1 string,p2 string)

INSERT OVERWRITE TABLE table_with_partition
partition (p1='a',p2)
SELECT 'blarr3'

In the above example, we do not have enough input columns for dynamic partitioning. The first input column is already taken as data columns. This PR is to issue an exception in this scenario.

For your reference, below is the behavior of Hive:

hive> insert overwrite table t9 partition(p1='1',p2) select col from t1;
FAILED: SemanticException [Error 10044]: Line 1:23 Cannot insert into target table because column number/types are different 'p2': Table insclause-0 has 2 columns, but query has 1 columns.

How was this patch tested?

Added a test case and fixed an existing test case

[gatorsmile]

There is a better fix in #12313. Let me close it.

[SparkQA]

Test build #60373 has finished for PR 13628 at commit 4c49112.

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