[
https://issues.apache.org/jira/browse/DRILL-6910?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Volodymyr Vysotskyi updated DRILL-6910:
---------------------------------------
Description:
*Data:*
{code:sql}
create table dfs.tmp.`nation` as select * from cp.`tpch/nation.parquet` where
n_regionkey < 10;
{code}
*Query:*
{code:sql}
explain plan for select n_nationkey from dfs.tmp.`nation` where n_regionkey <
10;
{code}
*Expected result:*
The filtering column (n_regionkey) should not be present in scan operator.
*Actual result:*
It remains in scan in spite of filter pruning.
{noformat}
00-00 Screen : rowType = RecordType(ANY n_nationkey): rowcount = 25.0,
cumulative cost = {52.5 rows, 77.5 cpu, 50.0 io, 0.0 network, 0.0 memory}, id =
112988
00-01 Project(n_nationkey=[$1]) : rowType = RecordType(ANY n_nationkey):
rowcount = 25.0, cumulative cost = {50.0 rows, 75.0 cpu, 50.0 io, 0.0 network,
0.0 memory}, id = 112987
00-02 Scan(table=[[dfs, tmp, nation]], groupscan=[ParquetGroupScan
[entries=[ReadEntryWithPath [path=maprfs:///tmp/nation]],
selectionRoot=maprfs:/tmp/nation, numFiles=1, numRowGroups=1,
usedMetadataFile=false, columns=[`n_regionkey`, `n_nationkey`]]]) : rowType =
RecordType(ANY n_regionkey, ANY n_nationkey): rowcount = 25.0, cumulative cost
= {25.0 rows, 50.0 cpu, 50.0 io, 0.0 network, 0.0 memory}, id = 112986
{noformat}
was:
*Data:*
{code:sql}
create table dfs.tmp.`nation` as select * from cp.`tpch/nation.parquet`;
{code}
*Query:*
{code:sql}
explain plan for select n_nationkey from dfs.tmp.`nation` where n_regionkey <
10;
{code}
*Expected result:*
The filtering column (n_regionkey) should not be present in scan operator.
*Actual result:*
It remains in scan in spite of filter pruning.
{noformat}
00-00 Screen : rowType = RecordType(ANY n_nationkey): rowcount = 25.0,
cumulative cost = {52.5 rows, 77.5 cpu, 50.0 io, 0.0 network, 0.0 memory}, id =
112988
00-01 Project(n_nationkey=[$1]) : rowType = RecordType(ANY n_nationkey):
rowcount = 25.0, cumulative cost = {50.0 rows, 75.0 cpu, 50.0 io, 0.0 network,
0.0 memory}, id = 112987
00-02 Scan(table=[[dfs, tmp, nation]], groupscan=[ParquetGroupScan
[entries=[ReadEntryWithPath [path=maprfs:///tmp/nation]],
selectionRoot=maprfs:/tmp/nation, numFiles=1, numRowGroups=1,
usedMetadataFile=false, columns=[`n_regionkey`, `n_nationkey`]]]) : rowType =
RecordType(ANY n_regionkey, ANY n_nationkey): rowcount = 25.0, cumulative cost
= {25.0 rows, 50.0 cpu, 50.0 io, 0.0 network, 0.0 memory}, id = 112986
{noformat}
> A filtering column remains in scan when filter pruning happens
> --------------------------------------------------------------
>
> Key: DRILL-6910
> URL: https://issues.apache.org/jira/browse/DRILL-6910
> Project: Apache Drill
> Issue Type: Improvement
> Affects Versions: 1.14.0
> Reporter: Anton Gozhiy
> Assignee: Volodymyr Vysotskyi
> Priority: Major
> Fix For: 1.16.0
>
>
> *Data:*
> {code:sql}
> create table dfs.tmp.`nation` as select * from cp.`tpch/nation.parquet` where
> n_regionkey < 10;
> {code}
> *Query:*
> {code:sql}
> explain plan for select n_nationkey from dfs.tmp.`nation` where n_regionkey <
> 10;
> {code}
> *Expected result:*
> The filtering column (n_regionkey) should not be present in scan operator.
> *Actual result:*
> It remains in scan in spite of filter pruning.
> {noformat}
> 00-00 Screen : rowType = RecordType(ANY n_nationkey): rowcount = 25.0,
> cumulative cost = {52.5 rows, 77.5 cpu, 50.0 io, 0.0 network, 0.0 memory}, id
> = 112988
> 00-01 Project(n_nationkey=[$1]) : rowType = RecordType(ANY n_nationkey):
> rowcount = 25.0, cumulative cost = {50.0 rows, 75.0 cpu, 50.0 io, 0.0
> network, 0.0 memory}, id = 112987
> 00-02 Scan(table=[[dfs, tmp, nation]], groupscan=[ParquetGroupScan
> [entries=[ReadEntryWithPath [path=maprfs:///tmp/nation]],
> selectionRoot=maprfs:/tmp/nation, numFiles=1, numRowGroups=1,
> usedMetadataFile=false, columns=[`n_regionkey`, `n_nationkey`]]]) : rowType =
> RecordType(ANY n_regionkey, ANY n_nationkey): rowcount = 25.0, cumulative
> cost = {25.0 rows, 50.0 cpu, 50.0 io, 0.0 network, 0.0 memory}, id = 112986
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
