[CARBONDATA-1282]Choose BatchedDatasource scan only if schema fits for codegen

[ashokblend]

Problem
When table is having large no of column say 2000, then spark gives code generation issue during full scan query as size of generated code exceeds 64KB.
Solution
As in code, we have two BatchedDataSourceScan and RowDataSourceScan to scan query. As per implementation BatchedDataSourceScan is used when code generation is supported else RowDataSourceScan . Spark checks the configuration spark.sql.codegen.wholeStage is enabled and also column size should not exceeds its configuration i.e spark.sql.codegen.maxFields. Hence we need to add one more check for spark.sql.codegen.maxFields.
Testing
Tested manually.

[asfgit]

Can one of the admins verify this patch?

[asfgit]

Can one of the admins verify this patch?

[CarbonDataQA]

Build Success with Spark 1.6, Please check CI http://144.76.159.231:8080/job/ApacheCarbonPRBuilder/372/

[CarbonDataQA]

Build Success with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder/2960/

[gvramana]

LGTM