[
https://issues.apache.org/jira/browse/SPARK-24020?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Petar Zecevic updated SPARK-24020:
----------------------------------
Description:
The problem we are solving is the case where you have two big tables
partitioned by X column, but also sorted within partitions by Y column and you
need to calculate an expensive function on the joined rows, which reduces the
number of output rows (e.g. condition based on a spatial distance calculation).
But you could theoretically reduce the number of joined rows for which the
calculation itself is performed by using a range condition on the Y column.
Something like this:
{{... WHERE t1.X = t2.X AND t1.Y BETWEEN t2.Y - d AND t2.Y + d AND <function
calculation...>}}
However, during a sort-merge join with this range condition specified, Spark
will first cross-join all the rows with the same X value and only then try to
apply the range condition and any function calculations. This happens because,
inside the generated sort-merge join (SMJ) code, these extra conditions are put
in the same block with the function being calculated and there is no way to
evaluate these conditions before reading all the rows to be checked into memory
(into an {{ExternalAppendOnlyUnsafeRowArray}}). If the two tables have a large
number of rows per X, this can result in a huge number of calculations and a
huge number of rows in executor memory, which can be unfeasible.
h2. The solution implementation
We therefore propose a change to the sort-merge join so that, when these extra
conditions are specified, a queue is used instead of the
ExternalAppendOnlyUnsafeRowArray class. This queue is then used as a moving
window through the values from the right relation as the left row changes. You
could call this a combination of an equi-join and a theta join; in literature
it is sometimes called an “epsilon join”. We call it a "sort-merge inner range
join".
This design uses much less memory (not all rows with the same values of X need
to be loaded into memory at once) and requires a much lower number of
comparisons (the validity of this statement depends on the actual data and
conditions used).
h3. The classes that need to be changed
For implementing the described change we propose changes to these classes:
* _ExtractEquiJoinKeys_ – a pattern that needs to be extended to be able to
recognize the case where a simple range condition with lower and upper limits
is used on a secondary column (a column not included in the equi-join
condition). The pattern also needs to extract the information later required
for code generation etc.
* _InMemoryUnsafeRowQueue_ – the moving window implementation to be used
instead of the _ExternalAppendOnlyUnsafeRowArray_ class. The rows need to be
removed and added to/from the structure as the left key (X) changes, or the
left secondary value (Y) changes, so the structure needs to be a queue. To make
the change as less intrusive as possible, we propose to implement
_InMemoryUnsafeRowQueue_ as a subclass of _ExternalAppendOnlyUnsafeRowArray_
* _JoinSelection_ – a strategy that uses _ExtractEquiJoinKeys_ and needs to be
aware of the extracted range conditions
* _SortMergeJoinExec_ – the main implementation of the optimization. Needs to
support two code paths:
** when whole-stage code generation is turned off (method doExecute, which uses
sortMergeJoinInnerRangeScanner)
** when whole-stage code generation is turned on (methods doProduce and
genScanner)
* _SortMergeJoinInnerRangeScanner_ – implements the SMJ with inner-range
optimization in the case when whole-stage codegen is turned off
* _InnerJoinSuite_ – functional tests
* _JoinBenchmark_ – performance tests
h2. Triggering the optimization
The optimization should be triggered automatically when an equi-join expression
is present AND lower and upper range conditions on a secondary column are
specified. If the tables aren't sorted by both columns, appropriate sorts
should be added.
To limit the impact of this change we also propose adding a new parameter
(tentatively named "spark.sql.join.smj.useInnerRangeOptimization") which could
be used to switch off the optimization entirely.
h2. Applicable use cases
Potential use-cases for this are joins based on spatial or temporal distance
calculations.
was:
The problem we are solving is the case where you have two big tables
partitioned by X column, but also sorted within partitions by Y column and you
need to calculate an expensive function on the joined rows, which reduces the
number of output rows (e.g. condition based on a spatial distance calculation).
But you could theoretically reduce the number of joined rows for which the
calculation itself is performed by using a range condition on the Y column.
Something like this:
{{... WHERE t1.X = t2.X AND t1.Y BETWEEN t2.Y - d AND t2.Y + d AND <function
calculation...>}}
However, during a sort-merge join with this range condition specified, Spark
will first cross-join all the rows with the same X value and only then try to
apply the range condition and any function calculations. This happens because,
inside the generated sort-merge join (SMJ) code, these extra conditions are put
in the same block with the function being calculated and there is no way to
evaluate these conditions before reading all the rows to be checked into memory
(into an {{ExternalAppendOnlyUnsafeRowArray}}). If the two tables have a large
number of rows per X, this can result in a huge number of calculations and a
huge number of rows in executor memory, which can be unfeasible.
h2. The solution implementation
We therefore propose a change to the sort-merge join so that, when these extra
conditions are specified, a queue is used instead of the
ExternalAppendOnlyUnsafeRowArray class. This queue is then used as a moving
window through the values from the right relation as the left row changes. You
could call this a combination of an equi-join and a theta join; in literature
it is sometimes called an “epsilon join”. We call it a "sort-merge inner range
join".
This design uses much less memory (not all rows with the same values of X need
to be loaded into memory at once) and requires a much lower number of
comparisons (the validity of this statement depends on the actual data and
conditions used).
h3.
For implementing the described change we propose changes to these classes:
* _ExtractEquiJoinKeys_ – a pattern that needs to be extended to be able to
recognize the case where a simple range condition with lower and upper limits
is used on a secondary column (a column not included in the equi-join
condition). The pattern also needs to extract the information later required
for code generation etc.
* _InMemoryUnsafeRowQueue_ – the moving window implementation to be used
instead of the _ExternalAppendOnlyUnsafeRowArray_ class. The rows need to be
removed and added to/from the structure as the left key (X) changes, or the
left secondary value (Y) changes, so the structure needs to be a queue. To make
the change as less intrusive as possible, we propose to implement
_InMemoryUnsafeRowQueue_ as a subclass of _ExternalAppendOnlyUnsafeRowArray_
* _JoinSelection_ – a strategy that uses _ExtractEquiJoinKeys_ and needs to be
aware of the extracted range conditions
* _SortMergeJoinExec_ – the main implementation of the optimization. Needs to
support two code paths:
** when whole-stage code generation is turned off (method doExecute, which uses
sortMergeJoinInnerRangeScanner)
** when whole-stage code generation is turned on (methods doProduce and
genScanner)
* _SortMergeJoinInnerRangeScanner_ – implements the SMJ with inner-range
optimization in the case when whole-stage codegen is turned off
* _InnerJoinSuite_ – functional tests
* _JoinBenchmark_ – performance tests
The optimization should be triggered automatically when an equi-join expression
is present AND lower and upper range conditions on a secondary column are
specified. If the tables aren't sorted by both columns, appropriate sorts
should be added.
To limit the impact of this change we also propose adding a new parameter
(tentatively named "spark.sql.join.smj.useInnerRangeOptimization") which could
be used to switch off the optimization entirely.
Potential use-cases for this are joins based on spatial or temporal distance
calculations.
> Sort-merge join inner range optimization
> ----------------------------------------
>
> Key: SPARK-24020
> URL: https://issues.apache.org/jira/browse/SPARK-24020
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 2.3.0
> Reporter: Petar Zecevic
> Priority: Major
> Attachments: SMJ-innerRange-PR24020-designDoc.pdf
>
>
> The problem we are solving is the case where you have two big tables
> partitioned by X column, but also sorted within partitions by Y column and
> you need to calculate an expensive function on the joined rows, which reduces
> the number of output rows (e.g. condition based on a spatial distance
> calculation). But you could theoretically reduce the number of joined rows
> for which the calculation itself is performed by using a range condition on
> the Y column. Something like this:
> {{... WHERE t1.X = t2.X AND t1.Y BETWEEN t2.Y - d AND t2.Y + d AND <function
> calculation...>}}
> However, during a sort-merge join with this range condition specified, Spark
> will first cross-join all the rows with the same X value and only then try to
> apply the range condition and any function calculations. This happens
> because, inside the generated sort-merge join (SMJ) code, these extra
> conditions are put in the same block with the function being calculated and
> there is no way to evaluate these conditions before reading all the rows to
> be checked into memory (into an {{ExternalAppendOnlyUnsafeRowArray}}). If the
> two tables have a large number of rows per X, this can result in a huge
> number of calculations and a huge number of rows in executor memory, which
> can be unfeasible.
> h2. The solution implementation
> We therefore propose a change to the sort-merge join so that, when these
> extra conditions are specified, a queue is used instead of the
> ExternalAppendOnlyUnsafeRowArray class. This queue is then used as a moving
> window through the values from the right relation as the left row changes.
> You could call this a combination of an equi-join and a theta join; in
> literature it is sometimes called an “epsilon join”. We call it a "sort-merge
> inner range join".
> This design uses much less memory (not all rows with the same values of X
> need to be loaded into memory at once) and requires a much lower number of
> comparisons (the validity of this statement depends on the actual data and
> conditions used).
> h3. The classes that need to be changed
> For implementing the described change we propose changes to these classes:
> * _ExtractEquiJoinKeys_ – a pattern that needs to be extended to be able to
> recognize the case where a simple range condition with lower and upper limits
> is used on a secondary column (a column not included in the equi-join
> condition). The pattern also needs to extract the information later required
> for code generation etc.
> * _InMemoryUnsafeRowQueue_ – the moving window implementation to be used
> instead of the _ExternalAppendOnlyUnsafeRowArray_ class. The rows need to be
> removed and added to/from the structure as the left key (X) changes, or the
> left secondary value (Y) changes, so the structure needs to be a queue. To
> make the change as less intrusive as possible, we propose to implement
> _InMemoryUnsafeRowQueue_ as a subclass of _ExternalAppendOnlyUnsafeRowArray_
> * _JoinSelection_ – a strategy that uses _ExtractEquiJoinKeys_ and needs to
> be aware of the extracted range conditions
> * _SortMergeJoinExec_ – the main implementation of the optimization. Needs to
> support two code paths:
> ** when whole-stage code generation is turned off (method doExecute, which
> uses sortMergeJoinInnerRangeScanner)
> ** when whole-stage code generation is turned on (methods doProduce and
> genScanner)
> * _SortMergeJoinInnerRangeScanner_ – implements the SMJ with inner-range
> optimization in the case when whole-stage codegen is turned off
> * _InnerJoinSuite_ – functional tests
> * _JoinBenchmark_ – performance tests
> h2. Triggering the optimization
> The optimization should be triggered automatically when an equi-join
> expression is present AND lower and upper range conditions on a secondary
> column are specified. If the tables aren't sorted by both columns,
> appropriate sorts should be added.
> To limit the impact of this change we also propose adding a new parameter
> (tentatively named "spark.sql.join.smj.useInnerRangeOptimization") which
> could be used to switch off the optimization entirely.
> h2. Applicable use cases
> Potential use-cases for this are joins based on spatial or temporal distance
> calculations.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
