You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "godfrey he (JIRA)" <ji...@apache.org> on 2019/04/15 08:42:00 UTC
[jira] [Updated] (FLINK-12192) Add support for generating optimized
logical plan for grouping sets and distinct aggregate
[ https://issues.apache.org/jira/browse/FLINK-12192?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
godfrey he updated FLINK-12192:
-------------------------------
Description:
This issue aims to supports generating optimized logical plan for grouping sets and distinct aggregate. (mentioned in FLINK-12076 and FLINK-12098)
for batch, query with distinct aggregate will be rewritten into two non-distinct aggregates by extended [AggregateExpandDistinctAggregatesRule|https://github.com/apache/calcite/blob/master/core/src/main/java/org/apache/calcite/rel/rules/AggregateExpandDistinctAggregatesRule.java], the first aggregate computes the distinct key and non-distinct aggregate function, and the second aggregate computes the distinct aggregate function based on first aggregate result. The first aggregate has grouping sets if there are more than one distinct aggregate functions on different fields.
for stream, query with distinct aggregate is handled by SplitAggregateRule in FLINK-12161.
query with grouping sets (or cube, rollup) will be rewritten into a regular aggregate with expand, and the expand node will duplicates the input data for each simple group.
e.g.
{noformat}
schema:
MyTable: a: INT, b: BIGINT, c: VARCHAR(32), d: VARCHAR(32)
Original records:
+-----+-----+-----+-----+
| a | b | c | d |
+-----+-----+-----+-----+
| 1 | 1 | c1 | d1 |
+-----+-----+-----+-----+
| 1 | 2 | c1 | d2 |
+-----+-----+-----+-----+
| 2 | 1 | c1 | d1 |
+-----+-----+-----+-----+
SELECT a, c, SUM(b) as b FROM MyTable GROUP BY GROUPING SETS (a, c)
logical plan after expanded:
LogicalCalc(expr#0..3=[{inputs}], proj#0..1=[{exprs}], b=[$t3])
LogicalAggregate(group=[{0, 2, 3}], groups=[[]], b=[SUM($1)])
LogicalExpand(projects=[{a=[$0], b=[$1], c=[null], $e=[1]}, {a=[null], b=[$1], c=[$2], $e=[2]}])
LogicalNativeTableScan(table=[[builtin, default, MyTable]])
notes:
'$e = 1' is equivalent to 'group by a'
'$e = 2' is equivalent to 'group by c'
expanded records:
+-----+-----+-----+-----+
| a | b | c | $e |
+-----+-----+-----+-----+ ---+---
| 1 | 1 | null| 1 | |
+-----+-----+-----+-----+ records expanded by record1
| null| 1 | c1 | 2 | |
+-----+-----+-----+-----+ ---+---
| 1 | 2 | null| 1 | |
+-----+-----+-----+-----+ records expanded by record2
| null| 2 | c1 | 2 | |
+-----+-----+-----+-----+ ---+---
| 2 | 1 | null| 1 | |
+-----+-----+-----+-----+ records expanded by record3
| null| 1 | c1 | 2 | |
+-----+-----+-----+-----+ ---+---
{noformat}
was:
This issue aims to supports generating optimized logical plan for grouping sets and distinct aggregate. (mentioned in [FLINK-12076|https://issues.apache.org/jira/browse/FLINK-12076] and [FLINK-12098|https://issues.apache.org/jira/browse/FLINK-12098])
for batch, query with distinct aggregate will be rewritten into two non-distinct aggregates by extended [AggregateExpandDistinctAggregatesRule|https://github.com/apache/calcite/blob/master/core/src/main/java/org/apache/calcite/rel/rules/AggregateExpandDistinctAggregatesRule.java], the first aggregate computes the distinct result and non-distinct aggregate function result, and the second aggregate computes the distinct aggregate function result based on first aggregate result. The first aggregate has grouping sets if there are more than one distinct aggregate on different fields.
for stream, query with distinct aggregate is handled by SplitAggregateRule in [FLINK-12161|https://issues.apache.org/jira/browse/FLINK-12161].
query with grouping sets (or cube, rollup) will be rewritten into a regular aggregate with expand.
The expand node will duplicates the input data for each simple group.
e.g.
{noformat}
schema:
MyTable: a: INT, b: BIGINT, c: VARCHAR(32), d: VARCHAR(32)
Original records:
+-----+-----+-----+-----+
| a | b | c | d |
+-----+-----+-----+-----+
| 1 | 1 | c1 | d1 |
+-----+-----+-----+-----+
| 1 | 2 | c1 | d2 |
+-----+-----+-----+-----+
| 2 | 1 | c1 | d1 |
+-----+-----+-----+-----+
SELECT a, c, SUM(b) as b FROM MyTable GROUP BY GROUPING SETS (a, c)
logical plan after expanded:
LogicalCalc(expr#0..3=[{inputs}], proj#0..1=[{exprs}], b=[$t3])
LogicalAggregate(group=[{0, 2, 3}], groups=[[]], b=[SUM($1)])
LogicalExpand(projects=[{a=[$0], b=[$1], c=[null], $e=[1]}, {a=[null], b=[$1], c=[$2], $e=[2]}])
LogicalNativeTableScan(table=[[builtin, default, MyTable]])
notes:
'$e = 1' is equivalent to 'group by a'
'$e = 2' is equivalent to 'group by c'
expanded records:
+-----+-----+-----+-----+
| a | b | c | $e |
+-----+-----+-----+-----+ ---+---
| 1 | 1 | null| 1 | |
+-----+-----+-----+-----+ records expanded by record1
| null| 1 | c1 | 2 | |
+-----+-----+-----+-----+ ---+---
| 1 | 2 | null| 1 | |
+-----+-----+-----+-----+ records expanded by record2
| null| 2 | c1 | 2 | |
+-----+-----+-----+-----+ ---+---
| 2 | 1 | null| 1 | |
+-----+-----+-----+-----+ records expanded by record3
| null| 1 | c1 | 2 | |
+-----+-----+-----+-----+ ---+---
{noformat}
> Add support for generating optimized logical plan for grouping sets and distinct aggregate
> ------------------------------------------------------------------------------------------
>
> Key: FLINK-12192
> URL: https://issues.apache.org/jira/browse/FLINK-12192
> Project: Flink
> Issue Type: New Feature
> Components: Table SQL / Planner
> Reporter: godfrey he
> Assignee: godfrey he
> Priority: Major
>
> This issue aims to supports generating optimized logical plan for grouping sets and distinct aggregate. (mentioned in FLINK-12076 and FLINK-12098)
> for batch, query with distinct aggregate will be rewritten into two non-distinct aggregates by extended [AggregateExpandDistinctAggregatesRule|https://github.com/apache/calcite/blob/master/core/src/main/java/org/apache/calcite/rel/rules/AggregateExpandDistinctAggregatesRule.java], the first aggregate computes the distinct key and non-distinct aggregate function, and the second aggregate computes the distinct aggregate function based on first aggregate result. The first aggregate has grouping sets if there are more than one distinct aggregate functions on different fields.
> for stream, query with distinct aggregate is handled by SplitAggregateRule in FLINK-12161.
> query with grouping sets (or cube, rollup) will be rewritten into a regular aggregate with expand, and the expand node will duplicates the input data for each simple group.
> e.g.
> {noformat}
> schema:
> MyTable: a: INT, b: BIGINT, c: VARCHAR(32), d: VARCHAR(32)
> Original records:
> +-----+-----+-----+-----+
> | a | b | c | d |
> +-----+-----+-----+-----+
> | 1 | 1 | c1 | d1 |
> +-----+-----+-----+-----+
> | 1 | 2 | c1 | d2 |
> +-----+-----+-----+-----+
> | 2 | 1 | c1 | d1 |
> +-----+-----+-----+-----+
> SELECT a, c, SUM(b) as b FROM MyTable GROUP BY GROUPING SETS (a, c)
> logical plan after expanded:
> LogicalCalc(expr#0..3=[{inputs}], proj#0..1=[{exprs}], b=[$t3])
> LogicalAggregate(group=[{0, 2, 3}], groups=[[]], b=[SUM($1)])
> LogicalExpand(projects=[{a=[$0], b=[$1], c=[null], $e=[1]}, {a=[null], b=[$1], c=[$2], $e=[2]}])
> LogicalNativeTableScan(table=[[builtin, default, MyTable]])
> notes:
> '$e = 1' is equivalent to 'group by a'
> '$e = 2' is equivalent to 'group by c'
> expanded records:
> +-----+-----+-----+-----+
> | a | b | c | $e |
> +-----+-----+-----+-----+ ---+---
> | 1 | 1 | null| 1 | |
> +-----+-----+-----+-----+ records expanded by record1
> | null| 1 | c1 | 2 | |
> +-----+-----+-----+-----+ ---+---
> | 1 | 2 | null| 1 | |
> +-----+-----+-----+-----+ records expanded by record2
> | null| 2 | c1 | 2 | |
> +-----+-----+-----+-----+ ---+---
> | 2 | 1 | null| 1 | |
> +-----+-----+-----+-----+ records expanded by record3
> | null| 1 | c1 | 2 | |
> +-----+-----+-----+-----+ ---+---
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)