[fix](nereids) pull up left join right predicate with or is null #58372
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
|
Thank you for your contribution to Apache Doris. Please clearly describe your PR:
|
feiniaofeiafei
force-pushed
the
fix_predicate
branch
from
6c2e7b9 to
3fdb1c7
Compare
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34257 ms |
TPC-DS: Total hot run time: 181245 ms |
ClickBench: Total hot run time: 27.86 s |
Contributor
FE Regression Coverage ReportIncrement line coverage |
feiniaofeiafei
force-pushed
the
fix_predicate
branch
2 times, most recently
from
a051974 to
7658f89
Compare
Contributor
Author
|
run buildall |
aytrack
reviewed
Nov 26, 2025
|
|
||
| // test left join right table predicate pull up | ||
| qt_leftjoin_right_pull_up_shape """ | ||
| explain shape plan select * from extend_infer_t3 t1 left join extend_infer_t4 t2 on t1.a=t2.a left join extend_infer_t5 t3 on t2.a= t3.a where t1.a=1; |
There was a problem hiding this comment.
add more cases：
- more t1 where condition:
t1.a is not null,t1.a in (1, 2) - more table join: t1 left join t2 left join t3 left join t4 ... and condition is the middle table
where t2.a = 1
Contributor
FE UT Coverage ReportIncrement line coverage |
TPC-H: Total hot run time: 34239 ms |
TPC-DS: Total hot run time: 183939 ms |
ClickBench: Total hot run time: 27.24 s |
Contributor
FE Regression Coverage ReportIncrement line coverage |
feiniaofeiafei
force-pushed
the
fix_predicate
branch
from
7658f89 to
1d1c12b
Compare
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34465 ms |
TPC-DS: Total hot run time: 183958 ms |
ClickBench: Total hot run time: 27.95 s |
Contributor
Author
|
run buildall |
TPC-H: Total hot run time: 34296 ms |
TPC-DS: Total hot run time: 182316 ms |
ClickBench: Total hot run time: 27.39 s |
Contributor
FE Regression Coverage ReportIncrement line coverage |
morrySnow
approved these changes
Dec 1, 2025
github-actions
bot
added
the
approved
Contributor
|
PR approved by at least one committer and no changes requested. |
Contributor
|
PR approved by anyone and no changes requested. |
englefly
approved these changes
Dec 1, 2025
feiniaofeiafei
added a commit
to feiniaofeiafei/doris
that referenced
this pull request
Dec 3, 2025
…che#58372) Related PR: apache#41731 Problem Summary: The optimizer cannot derive predicates across multiple LEFT JOINs. For example, given a filter on the leftmost table in a chain of LEFT JOINs, the optimizer should be able to derive predicates on the rightmost table, but it currently fails to do so. create table t1(a int, b int); create table t2(a int, b int); create table t3(a int, b int); insert into t1 values(1,2); insert into t2 values(1,2); insert into t3 values(1,2); insert into t3 values(null,2); explain logical plan select * from t1 left join t2 on t1.a=t2.a left join t3 on t2.a=t3.a where t1.a=1; LogicalResultSink[110] ( outputExprs=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalProject[109] ( distinct=false, projects=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalJoin[108] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#2 = a#4)], otherJoinConjuncts=[], markJoinConjuncts=[] ) |--LogicalProject[105] ( distinct=false, projects=[a#0, b#1, a#2, b#3] ) | +--LogicalJoin[104] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#0 = a#2)], otherJoinConjuncts=[], markJoinConjuncts=[] ) | |--LogicalFilter[101] ( predicates=(a#0 = 1) ) | | +--LogicalOlapScan ( qualified=internal.maldb.t1, indexName=<index_not_selected>, selectedIndexId=1764043369852, preAgg=ON, operativeCol=[a#0], virtualColumns=[] ) | +--LogicalFilter[103] ( predicates=(a#2 = 1) ) | +--LogicalOlapScan ( qualified=internal.maldb.t2, indexName=<index_not_selected>, selectedIndexId=1764043369875, preAgg=ON, operativeCol=[a#2], virtualColumns=[] ) +--LogicalOlapScan ( qualified=internal.maldb.t3, indexName=<index_not_selected>, selectedIndexId=1764043369898, preAgg=ON, operativeCol=[a#4], virtualColumns=[] ) The optimizer should derive t3.a=1 from t1.a=1 and the join conditions, but it currently doesn't. The root cause is that the PullUpPredicates rule doesn't properly handle predicate pull-up from the right side of LEFT JOINs. This PR fixes this by generating null-tolerant predicates when pulling up from RIGHT JOIN's right table and strengthening them when possible based on upper-level join conditions. after this pr: LogicalResultSink[110] ( outputExprs=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalProject[109] ( distinct=false, projects=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalJoin[108] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#2 = a#4)], otherJoinConjuncts=[], markJoinConjuncts=[] ) |--LogicalProject[105] ( distinct=false, projects=[a#0, b#1, a#2, b#3] ) | +--LogicalJoin[104] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#0 = a#2)], otherJoinConjuncts=[], markJoinConjuncts=[] ) | |--LogicalFilter[101] ( predicates=(a#0 = 1) ) | | +--LogicalOlapScan ( qualified=internal.maldb.t1, indexName=<index_not_selected>, selectedIndexId=1764043369852, preAgg=ON, operativeCol=[a#0], virtualColumns=[] ) | +--LogicalFilter[103] ( predicates=(a#2 = 1) ) | +--LogicalOlapScan ( qualified=internal.maldb.t2, indexName=<index_not_selected>, selectedIndexId=1764043369875, preAgg=ON, operativeCol=[a#2], virtualColumns=[] ) +--LogicalFilter[107] ( predicates=(a#4 = 1) ) +--LogicalOlapScan ( qualified=internal.maldb.t3, indexName=<index_not_selected>, selectedIndexId=1764043369898, preAgg=ON, operativeCol=[a#4], virtualColumns=[] )
feiniaofeiafei
added a commit
to feiniaofeiafei/doris
that referenced
this pull request
Dec 3, 2025
…che#58372) Related PR: apache#41731 Problem Summary: The optimizer cannot derive predicates across multiple LEFT JOINs. For example, given a filter on the leftmost table in a chain of LEFT JOINs, the optimizer should be able to derive predicates on the rightmost table, but it currently fails to do so. create table t1(a int, b int); create table t2(a int, b int); create table t3(a int, b int); insert into t1 values(1,2); insert into t2 values(1,2); insert into t3 values(1,2); insert into t3 values(null,2); explain logical plan select * from t1 left join t2 on t1.a=t2.a left join t3 on t2.a=t3.a where t1.a=1; LogicalResultSink[110] ( outputExprs=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalProject[109] ( distinct=false, projects=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalJoin[108] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#2 = a#4)], otherJoinConjuncts=[], markJoinConjuncts=[] ) |--LogicalProject[105] ( distinct=false, projects=[a#0, b#1, a#2, b#3] ) | +--LogicalJoin[104] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#0 = a#2)], otherJoinConjuncts=[], markJoinConjuncts=[] ) | |--LogicalFilter[101] ( predicates=(a#0 = 1) ) | | +--LogicalOlapScan ( qualified=internal.maldb.t1, indexName=<index_not_selected>, selectedIndexId=1764043369852, preAgg=ON, operativeCol=[a#0], virtualColumns=[] ) | +--LogicalFilter[103] ( predicates=(a#2 = 1) ) | +--LogicalOlapScan ( qualified=internal.maldb.t2, indexName=<index_not_selected>, selectedIndexId=1764043369875, preAgg=ON, operativeCol=[a#2], virtualColumns=[] ) +--LogicalOlapScan ( qualified=internal.maldb.t3, indexName=<index_not_selected>, selectedIndexId=1764043369898, preAgg=ON, operativeCol=[a#4], virtualColumns=[] ) The optimizer should derive t3.a=1 from t1.a=1 and the join conditions, but it currently doesn't. The root cause is that the PullUpPredicates rule doesn't properly handle predicate pull-up from the right side of LEFT JOINs. This PR fixes this by generating null-tolerant predicates when pulling up from RIGHT JOIN's right table and strengthening them when possible based on upper-level join conditions. after this pr: LogicalResultSink[110] ( outputExprs=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalProject[109] ( distinct=false, projects=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalJoin[108] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#2 = a#4)], otherJoinConjuncts=[], markJoinConjuncts=[] ) |--LogicalProject[105] ( distinct=false, projects=[a#0, b#1, a#2, b#3] ) | +--LogicalJoin[104] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#0 = a#2)], otherJoinConjuncts=[], markJoinConjuncts=[] ) | |--LogicalFilter[101] ( predicates=(a#0 = 1) ) | | +--LogicalOlapScan ( qualified=internal.maldb.t1, indexName=<index_not_selected>, selectedIndexId=1764043369852, preAgg=ON, operativeCol=[a#0], virtualColumns=[] ) | +--LogicalFilter[103] ( predicates=(a#2 = 1) ) | +--LogicalOlapScan ( qualified=internal.maldb.t2, indexName=<index_not_selected>, selectedIndexId=1764043369875, preAgg=ON, operativeCol=[a#2], virtualColumns=[] ) +--LogicalFilter[107] ( predicates=(a#4 = 1) ) +--LogicalOlapScan ( qualified=internal.maldb.t3, indexName=<index_not_selected>, selectedIndexId=1764043369898, preAgg=ON, operativeCol=[a#4], virtualColumns=[] )
nagisa-kunhah
pushed a commit
to nagisa-kunhah/doris
that referenced
this pull request
Dec 14, 2025
…che#58372) ### What problem does this PR solve? Related PR: apache#41731 Problem Summary: The optimizer cannot derive predicates across multiple LEFT JOINs. For example, given a filter on the leftmost table in a chain of LEFT JOINs, the optimizer should be able to derive predicates on the rightmost table, but it currently fails to do so. create table t1(a int, b int); create table t2(a int, b int); create table t3(a int, b int); insert into t1 values(1,2); insert into t2 values(1,2); insert into t3 values(1,2); insert into t3 values(null,2); explain logical plan select * from t1 left join t2 on t1.a=t2.a left join t3 on t2.a=t3.a where t1.a=1; LogicalResultSink[110] ( outputExprs=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalProject[109] ( distinct=false, projects=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalJoin[108] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#2 = a#4)], otherJoinConjuncts=[], markJoinConjuncts=[] ) |--LogicalProject[105] ( distinct=false, projects=[a#0, b#1, a#2, b#3] ) | +--LogicalJoin[104] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#0 = a#2)], otherJoinConjuncts=[], markJoinConjuncts=[] ) | |--LogicalFilter[101] ( predicates=(a#0 = 1) ) | | +--LogicalOlapScan ( qualified=internal.maldb.t1, indexName=<index_not_selected>, selectedIndexId=1764043369852, preAgg=ON, operativeCol=[a#0], virtualColumns=[] ) | +--LogicalFilter[103] ( predicates=(a#2 = 1) ) | +--LogicalOlapScan ( qualified=internal.maldb.t2, indexName=<index_not_selected>, selectedIndexId=1764043369875, preAgg=ON, operativeCol=[a#2], virtualColumns=[] ) +--LogicalOlapScan ( qualified=internal.maldb.t3, indexName=<index_not_selected>, selectedIndexId=1764043369898, preAgg=ON, operativeCol=[a#4], virtualColumns=[] ) The optimizer should derive t3.a=1 from t1.a=1 and the join conditions, but it currently doesn't. The root cause is that the PullUpPredicates rule doesn't properly handle predicate pull-up from the right side of LEFT JOINs. This PR fixes this by generating null-tolerant predicates when pulling up from RIGHT JOIN's right table and strengthening them when possible based on upper-level join conditions. after this pr: LogicalResultSink[110] ( outputExprs=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalProject[109] ( distinct=false, projects=[a#0, b#1, a#2, b#3, a#4, b#5] ) +--LogicalJoin[108] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#2 = a#4)], otherJoinConjuncts=[], markJoinConjuncts=[] ) |--LogicalProject[105] ( distinct=false, projects=[a#0, b#1, a#2, b#3] ) | +--LogicalJoin[104] ( type=LEFT_OUTER_JOIN, markJoinSlotReference=Optional.empty, hashJoinConjuncts=[(a#0 = a#2)], otherJoinConjuncts=[], markJoinConjuncts=[] ) | |--LogicalFilter[101] ( predicates=(a#0 = 1) ) | | +--LogicalOlapScan ( qualified=internal.maldb.t1, indexName=<index_not_selected>, selectedIndexId=1764043369852, preAgg=ON, operativeCol=[a#0], virtualColumns=[] ) | +--LogicalFilter[103] ( predicates=(a#2 = 1) ) | +--LogicalOlapScan ( qualified=internal.maldb.t2, indexName=<index_not_selected>, selectedIndexId=1764043369875, preAgg=ON, operativeCol=[a#2], virtualColumns=[] ) +--LogicalFilter[107] ( predicates=(a#4 = 1) ) +--LogicalOlapScan ( qualified=internal.maldb.t3, indexName=<index_not_selected>, selectedIndexId=1764043369898, preAgg=ON, operativeCol=[a#4], virtualColumns=[] )
w41ter
pushed a commit
to w41ter/incubator-doris
that referenced
this pull request
Dec 26, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What problem does this PR solve?
Issue Number: close #xxx
Related PR: #41731
Problem Summary:
Release note
The optimizer cannot derive predicates across multiple LEFT JOINs. For example, given a filter on the leftmost table in a chain of LEFT JOINs, the optimizer should be able to derive predicates on the rightmost table, but it currently fails to do so.
The optimizer should derive t3.a=1 from t1.a=1 and the join conditions, but it currently doesn't.
The root cause is that the PullUpPredicates rule doesn't properly handle predicate pull-up from the right side of LEFT JOINs. This PR fixes this by generating null-tolerant predicates when pulling up from RIGHT JOIN's right table and strengthening them when possible based on upper-level join conditions.
after this pr:
None
Check List (For Author)
Test
Behavior changed:
Does this need documentation?
Check List (For Reviewer who merge this PR)