[opt](Nereids) Optimize findValidItems method to handle circular dependencies #36839
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
|
Thank you for your contribution to Apache Doris. Since 2024-03-18, the Document has been moved to doris-website. |
|
run buildall |
TPC-H: Total hot run time: 40295 ms |
morrySnow
approved these changes
Jun 26, 2024
github-actions
bot
added
the
approved
|
PR approved by at least one committer and no changes requested. |
|
PR approved by anyone and no changes requested. |
TPC-DS: Total hot run time: 174048 ms |
ClickBench: Total hot run time: 30.45 s |
seawinde
approved these changes
Jun 26, 2024
dataroaring
pushed a commit
that referenced
this pull request
Jun 28, 2024
…ndencies (#36839) ## Proposed changes These optimizations allow the findValidItems method to correctly handle circular dependencies while maintaining the required output slots. The code is now more efficient and ensures that the necessary edges and items are preserved during the traversal process.
morrySnow
pushed a commit
that referenced
this pull request
Jul 5, 2024
…eliminate fail (#36888) this depends on #36839 #36886 Such as low level materialized view contains 5 group by dimension, and query also has 5 group by dimension, they are equals.In this scene, would not add aggregate on mv when try to rewrite query by materialized view. But if query only use 4 group by dimension and the remain demension is can be eliminated, then the query will change to 4 group by dimension. this will cause add aggregate on mv and will cause high level materialize rewrite fail later. Solution: in aggregate rewrite by materialized view, we try to eliminate mv group by dimension by query used dimension. if eliminate successfully. then high level will rewrite continue. such as low level mv def sql is as following: def join_mv_1 = """ select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, cast(sum(IFNULL(o_orderkey, 0) * IFNULL(o_custkey, 0)) as decimal(28, 8)) as agg1, sum(o_totalprice) as sum_total, max(o_totalprice) as max_total, min(o_totalprice) as min_total, count(*) as count_all, bitmap_union(to_bitmap(case when o_shippriority > 1 and o_orderkey IN (1, 3) then o_custkey else null end)) cnt_1, bitmap_union(to_bitmap(case when o_shippriority > 2 and o_orderkey IN (2) then o_custkey else null end)) as cnt_2 from lineitem_1 inner join orders_1 on lineitem_1.l_orderkey = orders_1.o_orderkey where lineitem_1.l_shipdate >= "2023-10-17" group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey """ def join_mv_2 = """ select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, t.agg1 as agg1, t.sum_total as agg3, t.max_total as agg4, t.min_total as agg5, t.count_all as agg6, cast(sum(IFNULL(ps_suppkey, 0) * IFNULL(ps_partkey, 0)) as decimal(28, 8)) as agg2 from ${mv_1} as t inner join partsupp_1 on t.l_partkey = partsupp_1.ps_partkey and t.l_suppkey = partsupp_1.ps_suppkey where partsupp_1.ps_suppkey > 1 group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, agg1, agg3, agg4, agg5, agg6 """ high level mv def sql is as following: def join_mv_3 = """ select t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, t2.agg1, >t1.agg2, t2.agg3, t1.agg4, t2.agg5, t1.agg6 from ${mv_2} as t1 left join ${mv_2} as t2 on t1.l_orderkey = t2.l_orderkey where t1.l_orderkey > 1 group by t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, >t2.agg1, >t1.agg2, t2.agg3, t1.agg4, t2.agg5, t1.agg6 """ if we run the query as following, it can hit the mv3 select t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, t2.agg1, >t1.agg2, >t2.agg3, t1.agg4, t2.agg5, t1.agg6 from ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, t.agg1 as agg1, t.sum_total as agg3, t.max_total as agg4, t.min_total as agg5, t.count_all as agg6, cast(sum(IFNULL(ps_suppkey, 0) * IFNULL(ps_partkey, 0)) as decimal(28, 8)) as agg2 from ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, cast(sum(IFNULL(o_orderkey, 0) * >IFNULL(o_custkey, 0)) as decimal(28, 8)) as agg1, sum(o_totalprice) as sum_total, max(o_totalprice) as max_total, min(o_totalprice) as min_total, count(*) as count_all, bitmap_union(to_bitmap(case when o_shippriority > 1 and o_orderkey IN (1, 3) then o_custkey else null end)) >cnt_1, bitmap_union(to_bitmap(case when o_shippriority > 2 and o_orderkey IN (2) then o_custkey else null end)) as >cnt_2 from lineitem_1 inner join orders_1 on lineitem_1.l_orderkey = orders_1.o_orderkey where lineitem_1.l_shipdate >= "2023-10-17" group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey ) as t inner join partsupp_1 on t.l_partkey = partsupp_1.ps_partkey and t.l_suppkey = partsupp_1.ps_suppkey where partsupp_1.ps_suppkey > 1 group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, agg1, agg3, agg4, >agg5, >agg6 ) as t1 left join ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, t.agg1 as agg1, t.sum_total as agg3, t.max_total as agg4, t.min_total as agg5, t.count_all as agg6, cast(sum(IFNULL(ps_suppkey, 0) * IFNULL(ps_partkey, 0)) as decimal(28, 8)) as agg2 from ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, cast(sum(IFNULL(o_orderkey, 0) * >IFNULL(o_custkey, 0)) as decimal(28, 8)) as agg1, sum(o_totalprice) as sum_total, max(o_totalprice) as max_total, min(o_totalprice) as min_total, count(*) as count_all, bitmap_union(to_bitmap(case when o_shippriority > 1 and o_orderkey IN (1, 3) then o_custkey else null end)) >cnt_1, bitmap_union(to_bitmap(case when o_shippriority > 2 and o_orderkey IN (2) then o_custkey else null end)) as >cnt_2 from lineitem_1 inner join orders_1 on lineitem_1.l_orderkey = orders_1.o_orderkey where lineitem_1.l_shipdate >= "2023-10-17" group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey ) as t inner join partsupp_1 on t.l_partkey = partsupp_1.ps_partkey and t.l_suppkey = partsupp_1.ps_suppkey where partsupp_1.ps_suppkey > 1 group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, agg1, agg3, agg4, agg5, >agg6 ) as t2 on t1.l_orderkey = t2.l_orderkey where t1.l_orderkey > 1 group by t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, >t2.agg1, >t1.agg2, t2.agg3, t1.agg4, t2.agg5, t1.agg6 --------- Co-authored-by: xiejiann <[email protected]>
dataroaring
pushed a commit
that referenced
this pull request
Jul 17, 2024
…eliminate fail (#36888) this depends on #36839 #36886 Such as low level materialized view contains 5 group by dimension, and query also has 5 group by dimension, they are equals.In this scene, would not add aggregate on mv when try to rewrite query by materialized view. But if query only use 4 group by dimension and the remain demension is can be eliminated, then the query will change to 4 group by dimension. this will cause add aggregate on mv and will cause high level materialize rewrite fail later. Solution: in aggregate rewrite by materialized view, we try to eliminate mv group by dimension by query used dimension. if eliminate successfully. then high level will rewrite continue. such as low level mv def sql is as following: def join_mv_1 = """ select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, cast(sum(IFNULL(o_orderkey, 0) * IFNULL(o_custkey, 0)) as decimal(28, 8)) as agg1, sum(o_totalprice) as sum_total, max(o_totalprice) as max_total, min(o_totalprice) as min_total, count(*) as count_all, bitmap_union(to_bitmap(case when o_shippriority > 1 and o_orderkey IN (1, 3) then o_custkey else null end)) cnt_1, bitmap_union(to_bitmap(case when o_shippriority > 2 and o_orderkey IN (2) then o_custkey else null end)) as cnt_2 from lineitem_1 inner join orders_1 on lineitem_1.l_orderkey = orders_1.o_orderkey where lineitem_1.l_shipdate >= "2023-10-17" group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey """ def join_mv_2 = """ select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, t.agg1 as agg1, t.sum_total as agg3, t.max_total as agg4, t.min_total as agg5, t.count_all as agg6, cast(sum(IFNULL(ps_suppkey, 0) * IFNULL(ps_partkey, 0)) as decimal(28, 8)) as agg2 from ${mv_1} as t inner join partsupp_1 on t.l_partkey = partsupp_1.ps_partkey and t.l_suppkey = partsupp_1.ps_suppkey where partsupp_1.ps_suppkey > 1 group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, agg1, agg3, agg4, agg5, agg6 """ high level mv def sql is as following: def join_mv_3 = """ select t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, t2.agg1, >t1.agg2, t2.agg3, t1.agg4, t2.agg5, t1.agg6 from ${mv_2} as t1 left join ${mv_2} as t2 on t1.l_orderkey = t2.l_orderkey where t1.l_orderkey > 1 group by t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, >t2.agg1, >t1.agg2, t2.agg3, t1.agg4, t2.agg5, t1.agg6 """ if we run the query as following, it can hit the mv3 select t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, t2.agg1, >t1.agg2, >t2.agg3, t1.agg4, t2.agg5, t1.agg6 from ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, t.agg1 as agg1, t.sum_total as agg3, t.max_total as agg4, t.min_total as agg5, t.count_all as agg6, cast(sum(IFNULL(ps_suppkey, 0) * IFNULL(ps_partkey, 0)) as decimal(28, 8)) as agg2 from ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, cast(sum(IFNULL(o_orderkey, 0) * >IFNULL(o_custkey, 0)) as decimal(28, 8)) as agg1, sum(o_totalprice) as sum_total, max(o_totalprice) as max_total, min(o_totalprice) as min_total, count(*) as count_all, bitmap_union(to_bitmap(case when o_shippriority > 1 and o_orderkey IN (1, 3) then o_custkey else null end)) >cnt_1, bitmap_union(to_bitmap(case when o_shippriority > 2 and o_orderkey IN (2) then o_custkey else null end)) as >cnt_2 from lineitem_1 inner join orders_1 on lineitem_1.l_orderkey = orders_1.o_orderkey where lineitem_1.l_shipdate >= "2023-10-17" group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey ) as t inner join partsupp_1 on t.l_partkey = partsupp_1.ps_partkey and t.l_suppkey = partsupp_1.ps_suppkey where partsupp_1.ps_suppkey > 1 group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, agg1, agg3, agg4, >agg5, >agg6 ) as t1 left join ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, t.agg1 as agg1, t.sum_total as agg3, t.max_total as agg4, t.min_total as agg5, t.count_all as agg6, cast(sum(IFNULL(ps_suppkey, 0) * IFNULL(ps_partkey, 0)) as decimal(28, 8)) as agg2 from ( select l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, cast(sum(IFNULL(o_orderkey, 0) * >IFNULL(o_custkey, 0)) as decimal(28, 8)) as agg1, sum(o_totalprice) as sum_total, max(o_totalprice) as max_total, min(o_totalprice) as min_total, count(*) as count_all, bitmap_union(to_bitmap(case when o_shippriority > 1 and o_orderkey IN (1, 3) then o_custkey else null end)) >cnt_1, bitmap_union(to_bitmap(case when o_shippriority > 2 and o_orderkey IN (2) then o_custkey else null end)) as >cnt_2 from lineitem_1 inner join orders_1 on lineitem_1.l_orderkey = orders_1.o_orderkey where lineitem_1.l_shipdate >= "2023-10-17" group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey ) as t inner join partsupp_1 on t.l_partkey = partsupp_1.ps_partkey and t.l_suppkey = partsupp_1.ps_suppkey where partsupp_1.ps_suppkey > 1 group by l_orderkey, l_partkey, l_suppkey, o_orderkey, o_custkey, ps_partkey, ps_suppkey, agg1, agg3, agg4, agg5, >agg6 ) as t2 on t1.l_orderkey = t2.l_orderkey where t1.l_orderkey > 1 group by t1.l_orderkey, t2.l_partkey, t1.l_suppkey, t2.o_orderkey, t1.o_custkey, t2.ps_partkey, t1.ps_suppkey, >t2.agg1, >t1.agg2, t2.agg3, t1.agg4, t2.agg5, t1.agg6 --------- Co-authored-by: xiejiann <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Proposed changes
These optimizations allow the findValidItems method to correctly handle circular dependencies while maintaining the required output slots. The code is now more efficient and ensures that the necessary edges and items are preserved during the traversal process.