Rahul Challapalli created DRILL-3613:
----------------------------------------
Summary: Partition Pruning not taking effect when we use a round
function and a cast on top of it
Key: DRILL-3613
URL: https://issues.apache.org/jira/browse/DRILL-3613
Project: Apache Drill
Issue Type: Bug
Components: Query Planning & Optimization
Affects Versions: 1.1.0
Reporter: Rahul Challapalli
Assignee: Jinfeng Ni
Fix For: 1.2.0
git.commit.id.abbrev=4d967ed
The below plan should only scan 2 files
{code}
explain plan for select columns[0], columns[1], columns[4], columns[10],
columns[13] from
`/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring`
where (dir0=cast(round(1993.11) as integer) and dir1='oct') or (dir0=1994 and
dir1='may');
+------+------+
| text | json |
+------+------+
| 00-00 Screen
00-01 Project(EXPR$0=[$0], EXPR$1=[$1], EXPR$2=[$2], EXPR$3=[$3],
EXPR$4=[$4])
00-02 Project(EXPR$0=[$2], EXPR$1=[$3], EXPR$2=[$4], EXPR$3=[$5],
EXPR$4=[$6])
00-03 SelectionVectorRemover
00-04 Filter(condition=[OR(AND(=($0, CAST(ROUND(1993.11)):INTEGER),
=($1, 'oct')), AND(=($0, 1994), =($1, 'may')))])
00-05 Project(dir0=[$0], dir1=[$2], ITEM=[ITEM($1, 0)],
ITEM3=[ITEM($1, 1)], ITEM4=[ITEM($1, 4)], ITEM5=[ITEM($1, 10)], ITEM6=[ITEM($1,
13)])
00-06 Scan(groupscan=[EasyGroupScan
[selectionRoot=maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring,
numFiles=8, columns=[`dir0`, `dir1`, `columns`[0], `columns`[1], `columns`[4],
`columns`[10], `columns`[13]],
files=[maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1991/oct/lineitemaj.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1992/oct/lineitemaj.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1996/oct/lineitemaj.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1994/may/lineitemae.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1994/oct/lineitemaj.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1993/oct/lineitemaj.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1997/oct/lineitemaj.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1995/oct/lineitemaj.tbl]]])
{code}
The above plan should only scan the below 2 files ideally
{code}
/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1994/may/lineitemae.tbl
/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1993/oct/lineitemaj.tbl
{code}
I tested with a different function below and the plan seems to be pruning
properly
{code}
explain plan for select columns[0], columns[1], columns[4], columns[10],
columns[13] from
`/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring`
where (dir0=cast(abs(-1993) as integer) and dir1='oct') or (dir0=1994 and
dir1='may');
+------+------+
| text | json |
+------+------+
| 00-00 Screen
00-01 Project(EXPR$0=[$0], EXPR$1=[$1], EXPR$2=[$2], EXPR$3=[$3],
EXPR$4=[$4])
00-02 Project(EXPR$0=[ITEM($1, 0)], EXPR$1=[ITEM($1, 1)],
EXPR$2=[ITEM($1, 4)], EXPR$3=[ITEM($1, 10)], EXPR$4=[ITEM($1, 13)])
00-03 Scan(groupscan=[EasyGroupScan
[selectionRoot=maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring,
numFiles=2, columns=[`dir0`, `dir1`, `columns`[0], `columns`[1], `columns`[4],
`columns`[10], `columns`[13]],
files=[maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1994/may/lineitemae.tbl,
maprfs:/drill/testdata/partition_pruning/hive/text/lineitem_hierarchical_intstring/1993/oct/lineitemaj.tbl]]])
{code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
