[SPARK-10849][SQL] Adding field metadata property to override default jdbc data source type mapping.

[sureshthalamati]

This patch allows users to override default type mapping of data frame field to database column type when writing data frame to jdbc data sources.

In some cases user may want to use specific database data type mapping for fields based on the database configuration (page size , type of table spaces ..etc) instead of the defaults. For example large varchar size for all the columns may not fit in row size limits , user may want to use mix of varchar , and clob types. Max precision supported in some database systems might be less than the spark decimal precision, in such cases user can use this option to adjust the decimal type precision , and scale to match the target database.

Added a new field meta data property name db.column.type . I am not sure what is the convention for these type of property names. Please let me know it it needs to be changed.

@rxin @marmbrus

[marmbrus]

ok to test

[SparkQA]

Test build #44586 has finished for PR 9352 at commit 4048c2d.

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

[sureshthalamati]

Failed test is test_trainOn_predictOn (main.StreamingKMeansTest). It seems to be unrelated to my changes. Can we retest this please.

[sureshthalamati]

Thinking about this more , I realized current version of the patch may introduce SQL injection. I will update the pull request with a new version of the fix.

[sureshthalamati]

Updated the patch to address sql injection issue by removing space characters from the input. Please review.

@marmbrus @rxin

[SparkQA]

Test build #44699 has finished for PR 9352 at commit ef26084.

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

[sureshthalamati]

retest this please.

Test failure is unrelated to my changes. Failed test is org.apache.spark.sql.sources.JsonHadoopFsRelationSuite.test all data types - TimestampType. It passes in my environment.

[rick-ibm]

Thanks for addressing the SQL injection concerns, Suresh. LGTM.

[sureshthalamati]

Jenkins, retest this please.

[tristanreid]

Anyone know the status of this change? Is there anything blocking, or was it superceded by something else? Thanks...

[AmplabJenkins]

Can one of the admins verify this patch?

[sureshthalamati]

Opened two new [WIP] PRs to fix this issue using different approaches.
#16208
#16209

Closing this PR.