[DRILL-7238] Drill does not use DirectScan for non-existent columns

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.17.0
• Component/s: Metadata
• Labels: ready-to-commit
• Assignee: Venkata Jyothsna Donapati
• Reporter: Venkata Jyothsna Donapati
• Created: 03/May/19 22:04
• Updated: 09/May/19 03:18
• Resolved: 09/May/19 03:18

Description

This query does not use the summary metadata cache file:
select count(int_nulls_id), count(int_id), count(ss_ticket_number), count(extra) from store_sales_null_blocks_int;

In this query, extra is a column that does not exist (non-existent column).

Here is the explain plan:

| 00-00    Screen
00-01      Project(EXPR$0=[$0], EXPR$1=[$1], EXPR$2=[$2], EXPR$3=[$3])
00-02        StreamAgg(group=[{}], EXPR$0=[$SUM0($0)], EXPR$1=[$SUM0($1)], EXPR$2=[$SUM0($2)], EXPR$3=[$SUM0($3)])
00-03          UnionExchange
01-01            StreamAgg(group=[{}], EXPR$0=[COUNT($0)], EXPR$1=[COUNT($1)], EXPR$2=[COUNT($2)], EXPR$3=[COUNT($3)])
01-02              Scan(table=[[dfs, parquet_metadata_cache, store_sales_null_blocks_int]], groupscan=[ParquetGroupScan [entries=[ReadEntryWithPath [path=/drill/testdata/metadata_cache/store_sales_null_blocks_int]], selectionRoot=/drill/testdata/metadata_cache/store_sales_null_blocks_int, numFiles=1, numRowGroups=11, usedMetadataFile=true, cacheFileRoot=/drill/testdata/metadata_cache/store_sales_null_blocks_int, columns=[`int_nulls_id`, `int_id`, `ss_ticket_number`, `extra`]]])

This is a regression from Drill 1.15.

Issue Links

• links to: GitHub Pull Request #1781