[Impala-ASF-CR] IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate
Impala Public Jenkins has posted comments on this change. ( http://gerrit.cloudera.org:8080/16720 ) Change subject: IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate .. Patch Set 43: Build Successful https://jenkins.impala.io/job/gerrit-code-review-checks/7910/ : Initial code review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun to run full precommit tests. -- To view, visit http://gerrit.cloudera.org:8080/16720 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I379405ee75b14929df7d6b5d20dabc6f51375691 Gerrit-Change-Number: 16720 Gerrit-PatchSet: 43 Gerrit-Owner: Qifan Chen Gerrit-Reviewer: Csaba Ringhofer Gerrit-Reviewer: Impala Public Jenkins Gerrit-Reviewer: Qifan Chen Gerrit-Reviewer: Tim Armstrong Gerrit-Reviewer: Zoltan Borok-Nagy Gerrit-Comment-Date: Sun, 27 Dec 2020 05:56:13 + Gerrit-HasComments: No
[Impala-ASF-CR] IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate
Impala Public Jenkins has posted comments on this change. ( http://gerrit.cloudera.org:8080/16720 ) Change subject: IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate .. Patch Set 42: Build Successful https://jenkins.impala.io/job/gerrit-code-review-checks/7909/ : Initial code review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun to run full precommit tests. -- To view, visit http://gerrit.cloudera.org:8080/16720 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I379405ee75b14929df7d6b5d20dabc6f51375691 Gerrit-Change-Number: 16720 Gerrit-PatchSet: 42 Gerrit-Owner: Qifan Chen Gerrit-Reviewer: Csaba Ringhofer Gerrit-Reviewer: Impala Public Jenkins Gerrit-Reviewer: Qifan Chen Gerrit-Reviewer: Tim Armstrong Gerrit-Reviewer: Zoltan Borok-Nagy Gerrit-Comment-Date: Sun, 27 Dec 2020 05:49:20 + Gerrit-HasComments: No
[Impala-ASF-CR] IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate
Qifan Chen has uploaded a new patch set (#43). ( http://gerrit.cloudera.org:8080/16720 ) Change subject: IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate .. IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate This patch adds a new class of predicates called overlap predicates to aid in the determination of whether a Parquet row group or a page overlap with a range computed from an equi hash join. If not, then the entire row group or page are skipped. An overlap predicate exists as a min/max filter. For the following query, the min and max in such a min/max filter are computed with the values from the join column from table 'b' and become fully available when the entire hash table is built. To evaluate the overlap predicate, these two values are compared against the min/max of each row group or page at the scan node for 'a'. select straight_join count(*) from lineitem_sorted_l_shipdate a join [SHUFFLE] lineitem_sorted_l_shipdate b where a.l_shipdate = b.l_receiptdate and b.l_commitdate = "1992-01-31"; An overlap predicate associated with the column type J (in hash table) and scan column type S will be formed when one of the following is true: Both J and S are booleans Both J and S are integers (tinyint, smallint, int, or bigint) Both J and S are approximate numeric (float or double) Both J and S are decimals with the same precision and scale Both J and S are strings (STRING, CHAR or VARCHAR) Both J and S are date Both J and S are timestamp Like any existing min/max filters, MAX_NUM_RUNTIME_FILTERS query option does not apply to min/max filters created for overlap predicates. The overlap predicates will always be evaluated, after the min/max conjuncts (if any). Two new run-time profile counters are added to report the number of row groups or pages filtered out via the overlap predicates respectively: 1. NumRuntimeFilteredRowGroups 2. NumRuntimeFilteredPages Testing: 1. Unit tested on various column types with TPCH and TPCDS tables. Benefits were significant when the join column on the outer table is sorted, or when the min/max boundary values of the pages or row groups are monotonic; 2. Added new tests in min_max_filters.test to demonstrate the number of filtered out pages and row groups with the two new profile counters; 2. Added new tests in runtime-filter-propagation.test to demonstrate that the overlap predicates work with different column types; 4. Added data type specific overlap method tests in min-max-filter-test.cc; 5. Core testing. TBD in this patch: 1. Performance measurement. To do in follow-up JIRAs: 1. Apply the overlap predicate on partition columns; 2. Apply the overlap predicate on each row; 3. IR code-gen for various MinMaxFilter::EvalOverlap methods. Change-Id: I379405ee75b14929df7d6b5d20dabc6f51375691 --- M be/src/exec/exec-node.h M be/src/exec/hdfs-scan-node-base.cc M be/src/exec/hdfs-scan-node-base.h M be/src/exec/parquet/hdfs-parquet-scanner.cc M be/src/exec/parquet/hdfs-parquet-scanner.h M be/src/exec/parquet/parquet-column-stats.cc M be/src/exec/parquet/parquet-column-stats.h M be/src/exec/partitioned-hash-join-builder.cc M be/src/exec/scan-node.cc M be/src/runtime/coordinator.cc M be/src/runtime/date-value.cc M be/src/runtime/date-value.h M be/src/runtime/runtime-filter-ir.cc M be/src/runtime/timestamp-value.cc M be/src/runtime/timestamp-value.h M be/src/service/query-options.cc M be/src/service/query-options.h M be/src/util/min-max-filter-test.cc M be/src/util/min-max-filter.cc M be/src/util/min-max-filter.h M common/thrift/ImpalaInternalService.thrift M common/thrift/ImpalaService.thrift M common/thrift/PlanNodes.thrift M fe/src/main/java/org/apache/impala/analysis/BinaryPredicate.java M fe/src/main/java/org/apache/impala/analysis/Predicate.java M fe/src/main/java/org/apache/impala/analysis/TupleDescriptor.java M fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java M fe/src/main/java/org/apache/impala/planner/RuntimeFilterGenerator.java M fe/src/test/java/org/apache/impala/planner/PlannerTest.java M testdata/workloads/functional-planner/queries/PlannerTest/aggregation.test M testdata/workloads/functional-planner/queries/PlannerTest/broadcast-bytes-limit-large.test M testdata/workloads/functional-planner/queries/PlannerTest/broadcast-bytes-limit.test A testdata/workloads/functional-planner/queries/PlannerTest/disable-runtime-overlap-filter.test M testdata/workloads/functional-planner/queries/PlannerTest/min-max-runtime-filters-hdfs-num-rows-est-enabled.test M testdata/workloads/functional-planner/queries/PlannerTest/min-max-runtime-filters.test M testdata/workloads/functional-planner/queries/PlannerTest/mt-dop-validation.test M testdata/workloads/functional-planner/queries/PlannerTest/nested-collections.test M
[Impala-ASF-CR] IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate
Impala Public Jenkins has posted comments on this change. ( http://gerrit.cloudera.org:8080/16720 ) Change subject: IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate .. Patch Set 42: (4 comments) http://gerrit.cloudera.org:8080/#/c/16720/42/fe/src/test/java/org/apache/impala/planner/PlannerTest.java File fe/src/test/java/org/apache/impala/planner/PlannerTest.java: http://gerrit.cloudera.org:8080/#/c/16720/42/fe/src/test/java/org/apache/impala/planner/PlannerTest.java@757 PS42, Line 757: options.setDisable_overlap_filter(true); // Required so that output doesn't vary by whether parquet tables are used or not. line too long (127 > 90) http://gerrit.cloudera.org:8080/#/c/16720/42/fe/src/test/java/org/apache/impala/planner/PlannerTest.java@787 PS42, Line 787: options.setDisable_overlap_filter(true); // Required so that output doesn't vary by the format of the table used. line has trailing whitespace http://gerrit.cloudera.org:8080/#/c/16720/42/fe/src/test/java/org/apache/impala/planner/PlannerTest.java@787 PS42, Line 787: options.setDisable_overlap_filter(true); // Required so that output doesn't vary by the format of the table used. line too long (118 > 90) http://gerrit.cloudera.org:8080/#/c/16720/42/tests/run-tests.py File tests/run-tests.py: http://gerrit.cloudera.org:8080/#/c/16720/42/tests/run-tests.py@219 PS42, Line 219: % flake8: E131 continuation line unaligned for hanging indent -- To view, visit http://gerrit.cloudera.org:8080/16720 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I379405ee75b14929df7d6b5d20dabc6f51375691 Gerrit-Change-Number: 16720 Gerrit-PatchSet: 42 Gerrit-Owner: Qifan Chen Gerrit-Reviewer: Csaba Ringhofer Gerrit-Reviewer: Impala Public Jenkins Gerrit-Reviewer: Qifan Chen Gerrit-Reviewer: Tim Armstrong Gerrit-Reviewer: Zoltan Borok-Nagy Gerrit-Comment-Date: Sun, 27 Dec 2020 05:28:09 + Gerrit-HasComments: Yes
[Impala-ASF-CR] IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate
Qifan Chen has uploaded a new patch set (#42). ( http://gerrit.cloudera.org:8080/16720 ) Change subject: IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate .. IMPALA-10325: Parquet scan should use min/max statistics to skip pages based on equi-join predicate This patch adds a new class of predicates called overlap predicates to aid in the determination of whether a Parquet row group or a page overlap with a range computed from an equi hash join. If not, then the entire row group or page are skipped. An overlap predicate exists as a min/max filter. For the following query, the min and max in such a min/max filter are computed with the values from the join column from table 'b' and become fully available when the entire hash table is built. To evaluate the overlap predicate, these two values are compared against the min/max of each row group or page at the scan node for 'a'. select straight_join count(*) from lineitem_sorted_l_shipdate a join [SHUFFLE] lineitem_sorted_l_shipdate b where a.l_shipdate = b.l_receiptdate and b.l_commitdate = "1992-01-31"; An overlap predicate associated with the column type J (in hash table) and scan column type S will be formed when one of the following is true: Both J and S are booleans Both J and S are integers (tinyint, smallint, int, or bigint) Both J and S are approximate numeric (float or double) Both J and S are decimals with the same precision and scale Both J and S are strings (STRING, CHAR or VARCHAR) Both J and S are date Both J and S are timestamp Like any existing min/max filters, MAX_NUM_RUNTIME_FILTERS query option does not apply to min/max filters created for overlap predicates. The overlap predicates will always be evaluated, after the min/max conjuncts (if any). Two new run-time profile counters are added to report the number of row groups or pages filtered out via the overlap predicates respectively: 1. NumRuntimeFilteredRowGroups 2. NumRuntimeFilteredPages Testing: 1. Unit tested on various column types with TPCH and TPCDS tables. Benefits were significant when the join column on the outer table is sorted, or when the min/max boundary values of the pages or row groups are monotonic; 2. Added new tests in min_max_filters.test to demonstrate the number of filtered out pages and row groups with the two new profile counters; 2. Added new tests in runtime-filter-propagation.test to demonstrate that the overlap predicates work with different column types; 4. Added data type specific overlap method tests in min-max-filter-test.cc; 5. Core testing. TBD in this patch: 1. Performance measurement. To do in follow-up JIRAs: 1. Apply the overlap predicate on partition columns; 2. Apply the overlap predicate on each row; 3. IR code-gen for various MinMaxFilter::EvalOverlap methods. Change-Id: I379405ee75b14929df7d6b5d20dabc6f51375691 --- M be/src/exec/exec-node.h M be/src/exec/hdfs-scan-node-base.cc M be/src/exec/hdfs-scan-node-base.h M be/src/exec/parquet/hdfs-parquet-scanner.cc M be/src/exec/parquet/hdfs-parquet-scanner.h M be/src/exec/parquet/parquet-column-stats.cc M be/src/exec/parquet/parquet-column-stats.h M be/src/exec/partitioned-hash-join-builder.cc M be/src/exec/scan-node.cc M be/src/runtime/coordinator.cc M be/src/runtime/date-value.cc M be/src/runtime/date-value.h M be/src/runtime/runtime-filter-ir.cc M be/src/runtime/timestamp-value.cc M be/src/runtime/timestamp-value.h M be/src/service/query-options.cc M be/src/service/query-options.h M be/src/util/min-max-filter-test.cc M be/src/util/min-max-filter.cc M be/src/util/min-max-filter.h M common/thrift/ImpalaInternalService.thrift M common/thrift/ImpalaService.thrift M common/thrift/PlanNodes.thrift M fe/src/main/java/org/apache/impala/analysis/BinaryPredicate.java M fe/src/main/java/org/apache/impala/analysis/Predicate.java M fe/src/main/java/org/apache/impala/analysis/TupleDescriptor.java M fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java M fe/src/main/java/org/apache/impala/planner/RuntimeFilterGenerator.java M fe/src/test/java/org/apache/impala/planner/PlannerTest.java M testdata/workloads/functional-planner/queries/PlannerTest/aggregation.test M testdata/workloads/functional-planner/queries/PlannerTest/broadcast-bytes-limit-large.test M testdata/workloads/functional-planner/queries/PlannerTest/broadcast-bytes-limit.test A testdata/workloads/functional-planner/queries/PlannerTest/disable-runtime-overlap-filter.test M testdata/workloads/functional-planner/queries/PlannerTest/min-max-runtime-filters-hdfs-num-rows-est-enabled.test M testdata/workloads/functional-planner/queries/PlannerTest/min-max-runtime-filters.test M testdata/workloads/functional-planner/queries/PlannerTest/mt-dop-validation.test M testdata/workloads/functional-planner/queries/PlannerTest/nested-collections.test M
[Impala-ASF-CR] IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu
Xianqing He has posted comments on this change. ( http://gerrit.cloudera.org:8080/16905 ) Change subject: IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu .. Patch Set 2: (1 comment) http://gerrit.cloudera.org:8080/#/c/16905/2/fe/src/main/java/org/apache/impala/analysis/SelectStmt.java File fe/src/main/java/org/apache/impala/analysis/SelectStmt.java: http://gerrit.cloudera.org:8080/#/c/16905/2/fe/src/main/java/org/apache/impala/analysis/SelectStmt.java@1130 PS2, Line 1130: // The predicates that can be bounded to KuduScanNode don't need to materialize > I think this is incorrect, only some conjuncts get pushed into Kudu (see ex I think it is not necessary to materialize the predicates pushed down to KuduScanNode here. Whether these predicates need to be materialized handled during the KuduScanNode#init. So if the prediacates can evaluate in kudu, it will not to be materilized. -- To view, visit http://gerrit.cloudera.org:8080/16905 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Iba8371eff6ae1bcffd51b44843175c52f2127e46 Gerrit-Change-Number: 16905 Gerrit-PatchSet: 2 Gerrit-Owner: Xianqing He Gerrit-Reviewer: Impala Public Jenkins Gerrit-Reviewer: Tim Armstrong Gerrit-Reviewer: Xianqing He Gerrit-Comment-Date: Sun, 27 Dec 2020 05:03:14 + Gerrit-HasComments: Yes
[Impala-ASF-CR] IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu
Tim Armstrong has posted comments on this change. ( http://gerrit.cloudera.org:8080/16905 ) Change subject: IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu .. Patch Set 2: (1 comment) http://gerrit.cloudera.org:8080/#/c/16905/2/fe/src/main/java/org/apache/impala/analysis/SelectStmt.java File fe/src/main/java/org/apache/impala/analysis/SelectStmt.java: http://gerrit.cloudera.org:8080/#/c/16905/2/fe/src/main/java/org/apache/impala/analysis/SelectStmt.java@1130 PS2, Line 1130: // The predicates that can be bounded to KuduScanNode don't need to materialize I think this is incorrect, only some conjuncts get pushed into Kudu (see extractKuduConjuncts()). For conjuncts that don't get pushed into Kudu, we still need to materialize the slots so we can evaluate them in the Impala KuduScanNode in the backend. Am I missing something? -- To view, visit http://gerrit.cloudera.org:8080/16905 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Iba8371eff6ae1bcffd51b44843175c52f2127e46 Gerrit-Change-Number: 16905 Gerrit-PatchSet: 2 Gerrit-Owner: Xianqing He Gerrit-Reviewer: Impala Public Jenkins Gerrit-Reviewer: Tim Armstrong Gerrit-Comment-Date: Sat, 26 Dec 2020 19:49:31 + Gerrit-HasComments: Yes
[Impala-ASF-CR] IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu
Impala Public Jenkins has posted comments on this change. ( http://gerrit.cloudera.org:8080/16905 ) Change subject: IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu .. Patch Set 2: Build Successful https://jenkins.impala.io/job/gerrit-code-review-checks/7908/ : Initial code review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun to run full precommit tests. -- To view, visit http://gerrit.cloudera.org:8080/16905 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Iba8371eff6ae1bcffd51b44843175c52f2127e46 Gerrit-Change-Number: 16905 Gerrit-PatchSet: 2 Gerrit-Owner: Xianqing He Gerrit-Reviewer: Impala Public Jenkins Gerrit-Comment-Date: Sat, 26 Dec 2020 16:15:42 + Gerrit-HasComments: No
[Impala-ASF-CR] IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu
Xianqing He has uploaded a new patch set (#2). ( http://gerrit.cloudera.org:8080/16905 ) Change subject: IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu .. IMPALA-10406: Query with analytic functions doesn't need to materialize the predicates bounded to kudu Before when query with analytic functions will materialize the unassigned conjuncts. But for the predicates that can be evaluated by kudu don't need to materialize. This optimization can reduce the amount of data to exchange and sort. Testing: - Add planner test in analytic-fns.test Change-Id: Iba8371eff6ae1bcffd51b44843175c52f2127e46 --- M fe/src/main/java/org/apache/impala/analysis/SelectStmt.java M testdata/workloads/functional-planner/queries/PlannerTest/analytic-fns.test 2 files changed, 135 insertions(+), 0 deletions(-) git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/05/16905/2 -- To view, visit http://gerrit.cloudera.org:8080/16905 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: newpatchset Gerrit-Change-Id: Iba8371eff6ae1bcffd51b44843175c52f2127e46 Gerrit-Change-Number: 16905 Gerrit-PatchSet: 2 Gerrit-Owner: Xianqing He Gerrit-Reviewer: Impala Public Jenkins
[Impala-ASF-CR] IMPALA-9922: A better approach to deal with date's sub-second fractions
fifteencai has posted comments on this change. ( http://gerrit.cloudera.org:8080/16869 ) Change subject: IMPALA-9922: A better approach to deal with date's sub-second fractions .. Patch Set 3: > Patch Set 3: > > (1 comment) Thank you so much, I am working on it -- To view, visit http://gerrit.cloudera.org:8080/16869 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I8e870bb8ad8fd14d388f37dfc5175589ecf9a5a7 Gerrit-Change-Number: 16869 Gerrit-PatchSet: 3 Gerrit-Owner: fifteencai Gerrit-Reviewer: Gabor Kaszab Gerrit-Reviewer: Impala Public Jenkins Gerrit-Reviewer: Tim Armstrong Gerrit-Reviewer: fifteencai Gerrit-Comment-Date: Sat, 26 Dec 2020 13:40:13 + Gerrit-HasComments: No