Details
-
Bug
-
Status: Resolved
-
Critical
-
Resolution: Fixed
-
None
-
None
-
ghx-label-13
Description
Let's say 'part_tbl' is a partitioned table. Its partition key is 'part_key'. Consider the following query:
select xxx from part_tbl where part_key=(select ... from dim_tbl);
Its query plan is a JoinNode with two ScanNodes. When estimating the cardinality of the JoinNode, the planner is not aware that 'part_key' is the partition column and the cardinality of the JoinNode should not be larger than the max row count across partitions.
The recent work in IMPALA-12018 (Consider runtime filter for cardinality reduction) helps in some cases since there are runtime filters on the partition column. But there are still some cases that we overestimate the cardinality. For instance, 'ss_sold_date_sk' is the only partition key of tpcds.store_sales. The following query
select count(*) from tpcds.store_sales where ss_sold_date_sk=( select min(d_date_sk) + 1000 from tpcds.date_dim);
has query plan:
+-------------------------------------------------------------+ | Explain String | +-------------------------------------------------------------+ | Max Per-Host Resource Reservation: Memory=18.94MB Threads=6 | | Per-Host Resource Estimates: Memory=243MB | | | | PLAN-ROOT SINK | | | | | 09:AGGREGATE [FINALIZE] | | | output: count:merge(*) | | | row-size=8B cardinality=1 | | | | | 08:EXCHANGE [UNPARTITIONED] | | | | | 04:AGGREGATE | | | output: count(*) | | | row-size=8B cardinality=1 | | | | | 03:HASH JOIN [LEFT SEMI JOIN, BROADCAST] | | | hash predicates: ss_sold_date_sk = min(d_date_sk) + 1000 | | | runtime filters: RF000 <- min(d_date_sk) + 1000 | | | row-size=4B cardinality=2.88M <---- Should be max(numRows) across partitions | | | | |--07:EXCHANGE [BROADCAST] | | | | | | | 06:AGGREGATE [FINALIZE] | | | | output: min:merge(d_date_sk) | | | | row-size=4B cardinality=1 | | | | | | | 05:EXCHANGE [UNPARTITIONED] | | | | | | | 02:AGGREGATE | | | | output: min(d_date_sk) | | | | row-size=4B cardinality=1 | | | | | | | 01:SCAN HDFS [tpcds.date_dim] | | | HDFS partitions=1/1 files=1 size=9.84MB | | | row-size=4B cardinality=73.05K | | | | | 00:SCAN HDFS [tpcds.store_sales] | | HDFS partitions=1824/1824 files=1824 size=346.60MB | | runtime filters: RF000 -> ss_sold_date_sk | | row-size=4B cardinality=2.88M | +-------------------------------------------------------------+
CC boroknagyz, rizaon
Attachments
Issue Links
- relates to
-
IMPALA-12018 Consider runtime filters in resource estimates
- Resolved