[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16102659#comment-16102659
]
Rui Li commented on HIVE-17087:
---
+1
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch, HIVE-17087.4.patch, HIVE-17087.5.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
> condition map:
>Inner Join 0 to 1
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16102338#comment-16102338
]
liyunzhang_intel commented on HIVE-17087:
-
[~stakiar]: GTM +1, meanwhile can you spend some time to review HIVE-16948,
another bug about dpp in HOS, thanks!
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch, HIVE-17087.4.patch, HIVE-17087.5.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
> condition map:
>Inner Join 0
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16101841#comment-16101841
]
Sahil Takiar commented on HIVE-17087:
-
[~lirui], [~kellyzly] patch has been update, any other comments?
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch, HIVE-17087.4.patch, HIVE-17087.5.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
> condition map:
>Inner Join 0 to 1
> keys:
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16100896#comment-16100896
]
Hive QA commented on HIVE-17087:
Here are the results of testing the latest attachment:
https://issues.apache.org/jira/secure/attachment/12878818/HIVE-17087.4.patch
{color:green}SUCCESS:{color} +1 due to 2 test(s) being added or modified.
{color:red}ERROR:{color} -1 due to 12 failed/errored test(s), 11103 tests
executed
*Failed tests:*
{noformat}
org.apache.hadoop.hive.cli.TestBeeLineDriver.testCliDriver[create_merge_compressed]
(batchId=240)
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver[auto_join6] (batchId=82)
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver[parquet_ppd_decimal]
(batchId=9)
org.apache.hadoop.hive.cli.TestMiniLlapCliDriver.testCliDriver[llap_smb]
(batchId=144)
org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[subquery_scalar]
(batchId=153)
org.apache.hadoop.hive.cli.TestMiniSparkOnYarnCliDriver.testCliDriver[spark_vectorized_dynamic_partition_pruning]
(batchId=168)
org.apache.hadoop.hive.cli.TestPerfCliDriver.testCliDriver[query14]
(batchId=235)
org.apache.hadoop.hive.cli.TestPerfCliDriver.testCliDriver[query23]
(batchId=235)
org.apache.hive.hcatalog.api.TestHCatClient.testPartitionRegistrationWithCustomSchema
(batchId=179)
org.apache.hive.hcatalog.api.TestHCatClient.testPartitionSpecRegistrationWithCustomSchema
(batchId=179)
org.apache.hive.hcatalog.api.TestHCatClient.testTableSchemaPropagation
(batchId=179)
org.apache.hive.jdbc.TestJdbcWithMiniHS2.testHttpRetryOnServerIdleTimeout
(batchId=228)
{noformat}
Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/6129/testReport
Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/6129/console
Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-6129/
Messages:
{noformat}
Executing org.apache.hive.ptest.execution.TestCheckPhase
Executing org.apache.hive.ptest.execution.PrepPhase
Executing org.apache.hive.ptest.execution.ExecutionPhase
Executing org.apache.hive.ptest.execution.ReportingPhase
Tests exited with: TestsFailedException: 12 tests failed
{noformat}
This message is automatically generated.
ATTACHMENT ID: 12878818 - PreCommit-HIVE-Build
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch, HIVE-17087.4.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16100209#comment-16100209
]
Sahil Takiar commented on HIVE-17087:
-
[~kellyzly] yes, I agree. Let's stick with the approach in the 3rd patch going
forward.
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch, HIVE-17087.4.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
> condition map:
>Inner Join 0 to 1
> keys:
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16099620#comment-16099620
]
liyunzhang_intel commented on HIVE-17087:
-
[~stakiar]: I think the 3rd patch is more clean to avoid the confusing order
problem mentioned in HIVE-10559.
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
> condition map:
>Inner Join 0 to 1
> keys:
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16099485#comment-16099485
]
Sahil Takiar commented on HIVE-17087:
-
Hey [~kellyzly]. Yes, opRules is a LinkedHashMap, so they should be run in
order. After looking at HIVE-10559 in more detail, I think there is a simpler
fix that can be made to avoid the NPE. I've left some comments in RB about it.
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch,
> HIVE-17087.3.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16098064#comment-16098064
]
liyunzhang_intel commented on HIVE-17087:
-
[~stakiar]
{quote}
It's actually suppose to be SparkMapJoinOptimizer and
SparkRemoveDynamicPruningBySize. I think the ordering may not be deterministic
because both are applied using the the same GraphWalker.
{quote}
opRules is LinkedHashMap which is insertion-ordered and why
SparkMapJoinOptimizer maybe executed later than SparkRemoveDynamicPruningBySize?
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16097765#comment-16097765
]
Sahil Takiar commented on HIVE-17087:
-
[~lirui], [~kellyzly], [~csun], [~janulatha] could you take a look at the
patch? https://reviews.apache.org/r/61044/ - let me know if you have any
questions, or if there is something that isn't clear.
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 2
> Stage: Stage-3
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: partitioned_table2
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col1 (type: int)
> 1 _col1 (type: int)
> Local Work:
> Map Reduce Local Work
> Stage: Stage-1
> Spark
> A masked pattern was here
> Vertices:
> Map 1
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Map Join Operator
> condition map:
>
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16096890#comment-16096890
]
Hive QA commented on HIVE-17087:
Here are the results of testing the latest attachment:
https://issues.apache.org/jira/secure/attachment/12878392/HIVE-17087.2.patch
{color:green}SUCCESS:{color} +1 due to 2 test(s) being added or modified.
{color:red}ERROR:{color} -1 due to 10 failed/errored test(s), 11093 tests
executed
*Failed tests:*
{noformat}
org.apache.hadoop.hive.cli.TestBeeLineDriver.testCliDriver[create_merge_compressed]
(batchId=240)
org.apache.hadoop.hive.cli.TestMiniLlapCliDriver.testCliDriver[llap_smb]
(batchId=144)
org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[columnstats_part_coltype]
(batchId=158)
org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[subquery_scalar]
(batchId=153)
org.apache.hadoop.hive.cli.TestMiniSparkOnYarnCliDriver.testCliDriver[spark_vectorized_dynamic_partition_pruning]
(batchId=168)
org.apache.hadoop.hive.cli.TestMiniTezCliDriver.testCliDriver[explainanalyze_3]
(batchId=99)
org.apache.hadoop.hive.cli.TestPerfCliDriver.testCliDriver[query14]
(batchId=235)
org.apache.hive.hcatalog.api.TestHCatClient.testPartitionRegistrationWithCustomSchema
(batchId=179)
org.apache.hive.hcatalog.api.TestHCatClient.testPartitionSpecRegistrationWithCustomSchema
(batchId=179)
org.apache.hive.hcatalog.api.TestHCatClient.testTableSchemaPropagation
(batchId=179)
{noformat}
Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/6109/testReport
Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/6109/console
Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-6109/
Messages:
{noformat}
Executing org.apache.hive.ptest.execution.TestCheckPhase
Executing org.apache.hive.ptest.execution.PrepPhase
Executing org.apache.hive.ptest.execution.ExecutionPhase
Executing org.apache.hive.ptest.execution.ReportingPhase
Tests exited with: TestsFailedException: 10 tests failed
{noformat}
This message is automatically generated.
ATTACHMENT ID: 12878392 - PreCommit-HIVE-Build
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch, HIVE-17087.2.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1, partitioned_table2 where
> partitioned_table1.part_col = partitioned_table2.part_col;
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-3 depends on stages: Stage-2
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 3
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16096479#comment-16096479
]
Sahil Takiar commented on HIVE-17087:
-
[~kellyzly] I'll update the comments to be a bit clearer. This was mostly
copied from {{OptimizeTezProcContext}}, specifically HIVE-10559. It's actually
suppose to be {{SparkMapJoinOptimizer}} and
{{SparkRemoveDynamicPruningBySize}}. I think the ordering may not be
deterministic because both are applied using the the same {{GraphWalker}}. I'm
not 100% sure though, but regardless I think the changes are still necessary to
avoid HIVE-10559.
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1 where partitioned_table1.part_col in
> (select regular_table.col from regular_table join partitioned_table2 on
> regular_table.col = partitioned_table2.part_col);
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-4 depends on stages: Stage-2
> Stage-5 depends on stages: Stage-4
> Stage-3 depends on stages: Stage-5
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 4
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 3
> Stage: Stage-4
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias: regular_table
> Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE
> Column stats: NONE
> Filter Operator
> predicate: col is not null (type: boolean)
> Statistics: Num rows: 1 Data size: 1 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 1 Data size: 1 Basic stats:
> COMPLETE Column stats: NONE
> Spark HashTable Sink Operator
> keys:
> 0 _col0 (type: int)
> 1 _col0 (type: int)
> Select Operator
> expressions: _col0 (type: int)
> outputColumnNames: _col0
> Statistics: Num
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16095886#comment-16095886
]
liyunzhang_intel commented on HIVE-17087:
-
[~stakiar]: 1 question about the patch
1.
{noformat}
/* Two of the optimization rules, ConvertJoinMapJoin and
RemoveDynamicPruningBySize, are put into
49 stats dependent optimizations and run together in TezCompiler.
There's no guarantee which one
50 runs first, but in either case, the prior one may have removed a
chain which the latter one is
51 not aware of. So we need to remember the leaf node(s) of that
chain so it can be skipped.
52
53 For example, as ConvertJoinMapJoin is removing the reduce sink, it
may also have removed a
54 dynamic partition pruning operator chain. However,
RemoveDynamicPruningBySize doesn't know this
55 and still tries to traverse that removed chain which will cause
NPE.
56
57 This may also happen when RemoveDynamicPruningBySize happens first.
58 */
59public HashSet
pruningOpsRemovedByPriorOpt;
{noformat}
[ConvertJoinMapJoin|https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/parse/TezCompiler.java#L155]
is inserted to opRules first and then
[RemoveDynamicPruningBySize|https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/parse/TezCompiler.java#L165]
and opRules is LinkedHashMap which is insertion-ordered, why say "There's no
guarantee which one runs first"?
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1 where partitioned_table1.part_col in
> (select regular_table.col from regular_table join partitioned_table2 on
> regular_table.col = partitioned_table2.part_col);
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-4 depends on stages: Stage-2
> Stage-5 depends on stages: Stage-4
> Stage-3 depends on stages: Stage-5
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 4
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Spark Partition Pruning Sink Operator
> partition key expr: part_col
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> target column name: part_col
> target work: Map 3
> Stage: Stage-4
> Spark
> A masked pattern was here
> Vertices:
> Map 2
> Map Operator Tree:
> TableScan
> alias:
[jira] [Commented] (HIVE-17087) Remove unnecessary HoS DPP trees during map-join conversion
[
https://issues.apache.org/jira/browse/HIVE-17087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16095710#comment-16095710
]
Hive QA commented on HIVE-17087:
Here are the results of testing the latest attachment:
https://issues.apache.org/jira/secure/attachment/12878290/HIVE-17087.1.patch
{color:green}SUCCESS:{color} +1 due to 2 test(s) being added or modified.
{color:red}ERROR:{color} -1 due to 11 failed/errored test(s), 11095 tests
executed
*Failed tests:*
{noformat}
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver[vector_row__id]
(batchId=46)
org.apache.hadoop.hive.cli.TestMiniLlapCliDriver.testCliDriver[llap_smb]
(batchId=144)
org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[subquery_scalar]
(batchId=153)
org.apache.hadoop.hive.cli.TestMiniSparkOnYarnCliDriver.testCliDriver[spark_dynamic_partition_pruning]
(batchId=168)
org.apache.hadoop.hive.cli.TestMiniSparkOnYarnCliDriver.testCliDriver[spark_vectorized_dynamic_partition_pruning]
(batchId=168)
org.apache.hadoop.hive.cli.TestPerfCliDriver.testCliDriver[query14]
(batchId=235)
org.apache.hadoop.hive.cli.TestPerfCliDriver.testCliDriver[query23]
(batchId=235)
org.apache.hive.hcatalog.api.TestHCatClient.testPartitionRegistrationWithCustomSchema
(batchId=179)
org.apache.hive.hcatalog.api.TestHCatClient.testPartitionSpecRegistrationWithCustomSchema
(batchId=179)
org.apache.hive.hcatalog.api.TestHCatClient.testTableSchemaPropagation
(batchId=179)
org.apache.hive.jdbc.TestJdbcWithMiniHS2.testHttpRetryOnServerIdleTimeout
(batchId=228)
{noformat}
Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/6103/testReport
Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/6103/console
Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-6103/
Messages:
{noformat}
Executing org.apache.hive.ptest.execution.TestCheckPhase
Executing org.apache.hive.ptest.execution.PrepPhase
Executing org.apache.hive.ptest.execution.ExecutionPhase
Executing org.apache.hive.ptest.execution.ReportingPhase
Tests exited with: TestsFailedException: 11 tests failed
{noformat}
This message is automatically generated.
ATTACHMENT ID: 12878290 - PreCommit-HIVE-Build
> Remove unnecessary HoS DPP trees during map-join conversion
> ---
>
> Key: HIVE-17087
> URL: https://issues.apache.org/jira/browse/HIVE-17087
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
>Reporter: Sahil Takiar
>Assignee: Sahil Takiar
> Attachments: HIVE-17087.1.patch
>
>
> Ran the following query in the {{TestSparkCliDriver}}:
> {code:sql}
> set hive.spark.dynamic.partition.pruning=true;
> set hive.auto.convert.join=true;
> create table partitioned_table1 (col int) partitioned by (part_col int);
> create table partitioned_table2 (col int) partitioned by (part_col int);
> create table regular_table (col int);
> insert into table regular_table values (1);
> alter table partitioned_table1 add partition (part_col = 1);
> insert into table partitioned_table1 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> alter table partitioned_table2 add partition (part_col = 1);
> insert into table partitioned_table2 partition (part_col = 1) values (1),
> (2), (3), (4), (5), (6), (7), (8), (9), (10);
> explain select * from partitioned_table1 where partitioned_table1.part_col in
> (select regular_table.col from regular_table join partitioned_table2 on
> regular_table.col = partitioned_table2.part_col);
> {code}
> and got the following explain plan:
> {code}
> STAGE DEPENDENCIES:
> Stage-2 is a root stage
> Stage-4 depends on stages: Stage-2
> Stage-5 depends on stages: Stage-4
> Stage-3 depends on stages: Stage-5
> Stage-1 depends on stages: Stage-3
> Stage-0 depends on stages: Stage-1
> STAGE PLANS:
> Stage: Stage-2
> Spark
> A masked pattern was here
> Vertices:
> Map 4
> Map Operator Tree:
> TableScan
> alias: partitioned_table1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: col (type: int), part_col (type: int)
> outputColumnNames: _col0, _col1
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Select Operator
> expressions: _col1 (type: int)
> outputColumnNames: _col0
> Statistics: Num rows: 10 Data size: 11 Basic stats:
> COMPLETE Column stats: NONE
> Group By Operator
> keys: _col0 (type: int)
> mode: hash
> outputColumnNames: _col0
>
