[DRILL-4106] Redundant Project on top of Scan in query plan

• Type: Bug
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Affects Version/s: 1.3.0
• Component/s: Execution - Relational Operators
• Assignee: Unassigned
• Reporter: Khurram Faraaz
• Created: 17/Nov/15 19:41
• Updated: 17/Nov/15 19:41

Description

Why doe we see two Projects after the Scan in the query plan ?
Table is auto partitioned by column c1
4 node cluster on CentOS, Drill 1.3, git.commit.id=a639c51c

#CTAS statement is,

CREATE TABLE inNstedDirAutoPrtn PARTITION BY(c1) AS SELECT cast(columns[0] AS INT) c1, cast(columns[1] AS BIGINT) c2, cast(columns[2] AS CHAR(2)) c3, cast(columns[3] AS VARCHAR(54)) c4, cast(columns[4] AS TIMESTAMP) c5, cast(columns[5] AS DATE) c6, cast(columns[6] as BOOLEAN) c7, cast(columns[7] as DOUBLE) c8, cast(columns[8] as TIME) c9 FROM `nested_dirs/data/csv/allData.csv`;

Why do we see two Projects on top of Scan in query plan ? One of them looks redundant.

0: jdbc:drill:schema=dfs.tmp> explain plan for select * from inNstedDirAutoPrtn where c1 IN (1,2,3,4,-1,0,100,-1710);
+------+------+
| text | json |
+------+------+
| 00-00    Screen
00-01      Project(*=[$0])
00-02        Project(*=[$0])
00-03          Scan(groupscan=[ParquetGroupScan [entries=[ReadEntryWithPath [path=/tmp/inNstedDirAutoPrtn/0_0_48.parquet], ReadEntryWithPath [path=/tmp/inNstedDirAutoPrtn/0_0_31.parquet], ReadEntryWithPath [path=/tmp/inNstedDirAutoPrtn/0_0_50.parquet], ReadEntryWithPath [path=/tmp/inNstedDirAutoPrtn/0_0_47.parquet], ReadEntryWithPath [path=/tmp/inNstedDirAutoPrtn/0_0_49.parquet], ReadEntryWithPath [path=/tmp/inNstedDirAutoPrtn/0_0_46.parquet]], selectionRoot=maprfs:/tmp/inNstedDirAutoPrtn, numFiles=6, usedMetadataFile=false, columns=[`*`]]])