Intelligent query processing in SQL databases

APPLIES TO: yesSQL Server yesAzure SQL Database noAzure SQL Data Warehouse noParallel Data Warehouse

The intelligent query processing (QP) feature family includes features with broad impact that improve the performance of existing workloads with minimal implementation effort to adopt.

Intelligent Query Processing

You can make workloads automatically eligible for intelligent query processing by enabling the applicable database compatibility level for the database. You can set this using Transact-SQL. For example:

ALTER DATABASE [WideWorldImportersDW] SET COMPATIBILITY_LEVEL = 150;

The following table details all intelligent query processing features, along with any requirement they have for database compatibility level.

IQP Feature Supported in Azure SQL Database Supported in SQL Server Description
Adaptive Joins (Batch Mode) Yes, under compatibility level 140 Yes, starting in SQL Server 2017 under compatibility level 140 Adaptive joins dynamically select a join type during runtime based on actual input rows.
Approximate Count Distinct Yes, public preview Yes, starting in SQL Server 2019 CTP 2.0, public preview Provide approximate COUNT DISTINCT for big data scenarios with the benefit of high performance and a low memory footprint.
Batch Mode on Rowstore Yes, under compatibility level 150, public preview Yes, starting in SQL Server 2019 CTP 2.0 under compatibility level 150, public preview Provide batch mode for CPU-bound relational DW workloads without requiring columnstore indexes.
Interleaved Execution Yes, under compatibility level 140 Yes, starting in SQL Server 2017 under compatibility level 140 Use the actual cardinality of the multi-statement table valued function encountered on first compilation instead of a fixed guess.
Memory Grant Feedback (Batch Mode) Yes, under compatibility level 140 Yes, starting in SQL Server 2017 under compatibility level 140 If a batch mode query has operations that spill to disk, add more memory for consecutive executions. If a query wastes > 50% of the memory allocated to it, reduce the memory grant side for consecutive executions.
Memory Grant Feedback (Row Mode) Yes, under compatibility level 150, public preview Yes, starting in SQL Server 2019 CTP 2.0 under compatibility level 150, public preview If a row mode query has operations that spill to disk, add more memory for consecutive executions. If a query wastes > 50% of the memory allocated to it, reduce the memory grant side for consecutive executions.
Scalar UDF Inlining No Yes, starting in SQL Server 2019 CTP 2.1 under compatibility level 150, public preview Scalar UDFs are transformed into equivalent relational expressions that are "inlined" into the calling query, often resulting in significant performance gains.
Table Variable Deferred Compilation Yes, under compatibility level 150, public preview Yes, starting in SQL Server 2019 CTP 2.0 under compatibility level 150, public preview Use the actual cardinality of the table variable encountered on first compilation instead of a fixed guess.

Batch mode adaptive joins

With this feature, your plan can dynamically switch to a better join strategy during execution by using a single cached plan.

The batch mode Adaptive Joins feature enables the choice of a Hash Join or Nested Loops Join method to be deferred until after the first input has been scanned. The Adaptive Join operator defines a threshold that is used to decide when to switch to a Nested Loops plan. Your plan can therefore dynamically switch to a better join strategy during execution. Here's how it works:

  • If the row count of the build join input is small enough that a nested loop join would be more optimal than a Hash Join, your plan switches to a Nested Loops algorithm.
  • If the build join input exceeds a specific row count threshold, no switch occurs and your plan continues with a Hash Join.

The following query is used to illustrate an Adaptive Join example:

SELECT [fo].[Order Key], [si].[Lead Time Days], [fo].[Quantity]
FROM [Fact].[Order] AS [fo]
INNER JOIN [Dimension].[Stock Item] AS [si]
       ON [fo].[Stock Item Key] = [si].[Stock Item Key]
WHERE [fo].[Quantity] = 360;

The query returns 336 rows. Enabling Live Query Statistics, we see the following plan:

Query result 336 rows

In the plan, we see the following:

  1. We have a columnstore index scan used to provide rows for the hash join build phase.
  2. We have the new Adaptive Join operator. This operator defines a threshold that is used to decide when to switch to a Nested Loops plan. For our example, the threshold is 78 rows. Anything with >= 78 rows will use a Hash Join. If less than the threshold, a Nested Loops Join will be used.
  3. Since we return 336 rows, we are exceeding the threshold and so the second branch represents the probe phase of a standard Hash Join operation. Notice that Live Query Statistics shows rows flowing through the operators - in this case "672 of 672".
  4. And the last branch is our Clustered Index Seek for use by the nested loop join had the threshold not been exceeded. Notice that we see "0 of 336" rows displayed (the branch is unused).

We are listening: If you find something outdated or incorrect in this article, such as a step or a code example, please tell us. You can click the This page button in the Feedback section at the bottom of this page. We read every item of feedback about SQL, typically the next day. Thanks.

Now contrast the plan with the same query, but this time for a Quantity value that only has one row in the table:

SELECT [fo].[Order Key], [si].[Lead Time Days], [fo].[Quantity]
FROM [Fact].[Order] AS [fo]
INNER JOIN [Dimension].[Stock Item] AS [si]
       ON [fo].[Stock Item Key] = [si].[Stock Item Key]
WHERE [fo].[Quantity] = 361;

The query returns one row. Enabling Live Query Statistics we see the following plan:

Query result one row

In the plan, we see the following:

  • With one row returned, the Clustered Index Seek now has rows flowing through it.
  • And since the Hash Join build phase did not continue, there are no rows flowing through the second branch.

Adaptive Join benefits

Workloads with frequent oscillations between small and large join input scans will benefit most from this feature.

Adaptive Join overhead

Adaptive joins introduce a higher memory requirement than an indexed Nested Loops Join equivalent plan. The additional memory is requested as if the Nested Loops was a Hash Join. There is also overhead for the build phase as a stop-and-go operation versus a Nested Loops streaming equivalent join. With that additional cost comes flexibility for scenarios where row counts may fluctuate in the build input.

Adaptive Join caching and re-use

Batch mode Adaptive Joins work for the initial execution of a statement, and once compiled, consecutive executions will remain adaptive based on the compiled Adaptive Join threshold and the runtime rows flowing through the build phase of the outer input.

Tracking Adaptive Join activity

The Adaptive Join operator has the following plan operator attributes:

Plan attribute Description
AdaptiveThresholdRows Shows the threshold use to switch from a hash join to nested loop join.
EstimatedJoinType What the join type is likely to be.
ActualJoinType In an actual plan, shows what join algorithm was ultimately chosen based on the threshold.

The estimated plan shows the Adaptive Join plan shape, along with a defined Adaptive Join threshold and estimated join type.

Adaptive join and Query Store interoperability

Query Store captures and is able to force a batch mode Adaptive Join plan.

Adaptive join eligible statements

A few conditions make a logical join eligible for a batch mode Adaptive Join:

  • The database compatibility level is 140.
  • The query is a SELECT statement (data modification statements are currently ineligible).
  • The join is eligible to be executed both by an indexed Nested Loops Join or a Hash Join physical algorithm.
  • The Hash Join uses batch mode - either through the presence of a Columnstore index in the query overall or a Columnstore indexed table being referenced directly by the join.
  • The generated alternative solutions of the Nested Loops Join and Hash Join should have the same first child (outer reference).

Adaptive joins and nested loop efficiency

If an Adaptive Join switches to a Nested Loops operation, it uses the rows already read by the Hash Join build. The operator does not re-read the outer reference rows again.

Adaptive threshold rows

The following chart shows an example intersection between the cost of a Hash Join versus the cost of a Nested Loops Join alternative. At this intersection point, the threshold is determined that in turn determines the actual algorithm used for the join operation.

Join threshold

Disabling adaptive joins without changing the compatibility level

Adaptive joins can be disabled at the database or statement scope while still maintaining database compatibility level 140 and higher.
To disable adaptive joins for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET DISABLE_BATCH_MODE_ADAPTIVE_JOINS = ON;

When enabled, this setting will appear as enabled in sys.database_scoped_configurations. To re-enable adaptive joins for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET DISABLE_BATCH_MODE_ADAPTIVE_JOINS = OFF;

You can also disable adaptive joins for a specific query by designating DISABLE_BATCH_MODE_ADAPTIVE_JOINS as a USE HINT query hint. For example:

SELECT s.CustomerID,
       s.CustomerName,
       sc.CustomerCategoryName
FROM Sales.Customers AS s
LEFT OUTER JOIN Sales.CustomerCategories AS sc
       ON s.CustomerCategoryID = sc.CustomerCategoryID
OPTION (USE HINT('DISABLE_BATCH_MODE_ADAPTIVE_JOINS')); 

A USE HINT query hint takes precedence over a database scoped configuration or trace flag setting.

Batch mode memory grant feedback

A query's post-execution plan in SQL Server includes the minimum required memory needed for execution and the ideal memory grant size to have all rows fit in memory. Performance suffers when memory grant sizes are incorrectly sized. Excessive grants result in wasted memory and reduced concurrency. Insufficient memory grants cause expensive spills to disk. By addressing repeating workloads, batch mode memory grant feedback recalculates the actual memory required for a query and then updates the grant value for the cached plan. When an identical query statement is executed, the query uses the revised memory grant size, reducing excessive memory grants that impact concurrency and fixing underestimated memory grants that cause expensive spills to disk. The following graph shows one example of using batch mode adaptive memory grant feedback. For the first execution of the query, duration was 88 seconds due to high spills:

DECLARE @EndTime datetime = '2016-09-22 00:00:00.000';
DECLARE @StartTime datetime = '2016-09-15 00:00:00.000';
SELECT TOP 10 hash_unique_bigint_id
FROM dbo.TelemetryDS
WHERE Timestamp BETWEEN @StartTime and @EndTime
GROUP BY hash_unique_bigint_id
ORDER BY MAX(max_elapsed_time_microsec) DESC;

High spills

With memory grant feedback enabled, for the second execution, duration is 1 second (down from 88 seconds), spills are removed entirely, and the grant is higher:

No spills

Memory grant feedback sizing

For an excessive memory grant condition, if the granted memory is more than two times the size of the actual used memory, memory grant feedback will recalculate the memory grant and update the cached plan. Plans with memory grants under 1 MB will not be recalculated for overages. For an insufficiently sized memory grant condition, that result in a spill to disk for batch mode operators, memory grant feedback will trigger a recalculation of the memory grant. Spill events are reported to memory grant feedback and can be surfaced via the spilling_report_to_memory_grant_feedback xEvent. This event returns the node id from the plan and spilled data size of that node.

Memory grant feedback and parameter sensitive scenarios

Different parameter values may also require different query plans in order to remain optimal. This type of query is defined as "parameter-sensitive." For parameter-sensitive plans, memory grant feedback will disable itself on a query if it has unstable memory requirements. The plan is disabled after several repeated runs of the query and this can be observed by monitoring the memory_grant_feedback_loop_disabled xEvent. For more information about parameter sniffing and parameter sensitivity, refer to the Query Processing Architecture Guide.

Memory grant feedback caching

Feedback can be stored in the cached plan for a single execution. It is the consecutive executions of that statement, however, that benefit from the memory grant feedback adjustments. This feature applies to repeated execution of statements. Memory grant feedback will change only the cached plan. Changes are currently not captured in the Query Store. Feedback is not persisted if the plan is evicted from cache. Feedback will also be lost if there is a failover. A statement using OPTION (RECOMPILE) creates a new plan and does not cache it. Since it is not cached, no memory grant feedback is produced and it is not stored for that compilation and execution. However, if an equivalent statement (that is, with the same query hash) that did not use OPTION (RECOMPILE) was cached and then re-executed, the consecutive statement can benefit from memory grant feedback.

Tracking memory grant feedback activity

You can track memory grant feedback events using the memory_grant_updated_by_feedback xEvent. This event tracks the current execution count history, the number of times the plan has been updated by memory grant feedback, the ideal additional memory grant before modification and the ideal additional memory grant after memory grant feedback has modified the cached plan.

Memory grant feedback, resource governor and query hints

The actual memory granted honors the query memory limit determined by the resource governor or query hint.

Disabling batch mode memory grant feedback without changing the compatibility level

Memory grant feedback can be disabled at the database or statement scope while still maintaining database compatibility level 140 and higher. To disable batch mode memory grant feedback for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET DISABLE_BATCH_MODE_MEMORY_GRANT_FEEDBACK = ON;

When enabled, this setting will appear as enabled in sys.database_scoped_configurations.

To re-enable batch mode memory grant feedback for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET DISABLE_BATCH_MODE_MEMORY_GRANT_FEEDBACK = OFF;

You can also disable batch mode memory grant feedback for a specific query by designating DISABLE_BATCH_MODE_MEMORY_GRANT_FEEDBACK as a USE HINT query hint. For example:

SELECT * FROM Person.Address  
WHERE City = 'SEATTLE' AND PostalCode = 98104
OPTION (USE HINT ('DISABLE_BATCH_MODE_MEMORY_GRANT_FEEDBACK')); 

A USE HINT query hint takes precedence over a database scoped configuration or trace flag setting.

Row mode memory grant feedback

Applies to: Azure SQL Database as a public preview feature

Note

Row mode memory grant feedback is a public preview feature.

Row mode memory grant feedback expands on the batch mode memory grant feedback feature by adjusting memory grant sizes for both batch and row mode operators.

To enable the public preview of row mode memory grant feedback in Azure SQL Database, enable database compatibility level 150 for the database you are connected to when executing the query.

Row mode memory grant feedback activity will be visible via the memory_grant_updated_by_feedback XEvent.

Starting with row mode memory grant feedback, two new query plan attributes will be shown for actual post-execution plans: IsMemoryGrantFeedbackAdjusted and LastRequestedMemory, which are added to the MemoryGrantInfo query plan XML element.

LastRequestedMemory shows the granted memory in Kilobytes (KB) from the prior query execution. IsMemoryGrantFeedbackAdjusted attribute allows you to check the state of memory grant feedback for the statement within an actual query execution plan. Values surfaced in this attribute are as follows:

IsMemoryGrantFeedbackAdjusted Value Description
No: First Execution Memory grant feedback does not adjust memory for the first compile and associated execution.
No: Accurate Grant If there is no spill to disk and the statement uses at least 50% of the granted memory, then memory grant feedback is not triggered.
No: Feedback disabled If memory grant feedback is continually triggered and fluctuates between memory-increase and memory-decrease operations, we will disable memory grant feedback for the statement.
Yes: Adjusting Memory grant feedback has been applied and may be further adjusted for the next execution.
Yes: Stable Memory grant feedback has been applied and granted memory is now stable, meaning that what was last granted for the previous execution is what was granted for the current execution.

Note

The public preview row mode memory grant feedback plan attributes are visible in SQL Server Management Studio graphical query execution plans in versions 17.9 and higher.

Disabling row mode memory grant feedback without changing the compatibility level

Row mode memory grant feedback can be disabled at the database or statement scope while still maintaining database compatibility level 150 and higher. To disable row mode memory grant feedback for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET ROW_MODE_MEMORY_GRANT_FEEDBACK = OFF;

To re-enable row mode memory grant feedback for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET ROW_MODE_MEMORY_GRANT_FEEDBACK = ON;

You can also disable row mode memory grant feedback for a specific query by designating DISABLE_ROW_MODE_MEMORY_GRANT_FEEDBACK as a USE HINT query hint. For example:

SELECT * FROM Person.Address  
WHERE City = 'SEATTLE' AND PostalCode = 98104
OPTION (USE HINT ('DISABLE_ROW_MODE_MEMORY_GRANT_FEEDBACK')); 

A USE HINT query hint takes precedence over a database scoped configuration or trace flag setting.

Interleaved execution for MSTVFs

With interleaved execution, the actual row counts from the function are used to make better-informed downstream query plan decisions. For more information on multi-statement table-valued functions (MSTVFs), see Table-valued functions.

Interleaved execution changes the unidirectional boundary between the optimization and execution phases for a single-query execution and enables plans to adapt based on the revised cardinality estimates. During optimization if we encounter a candidate for interleaved execution, which is currently multi-statement table-valued functions (MSTVFs), we will pause optimization, execute the applicable subtree, capture accurate cardinality estimates, and then resume optimization for downstream operations.

MSTVFs have a fixed cardinality guess of 100 starting with SQL Server 2014 (12.x), and 1 for earlier SQL Server versions. Interleaved execution helps workload performance issues that are due to these fixed cardinality estimates associated with MSTVFs. For more information on MSTVFs, see Create User-defined Functions (Database Engine).

The following image depicts a Live Query Statistics output, a subset of an overall execution plan that shows the impact of fixed cardinality estimates from MSTVFs. You can see the actual row flow vs. estimated rows. There are three noteworthy areas of the plan (flow is from right to left):

  1. The MSTVF Table Scan has a fixed estimate of 100 rows. For this example, however, there are 527,597 rows flowing through this MSTVF Table Scan, as seen in Live Query Statistics via the 527597 of 100 actual of estimated - so the fixed estimate is significantly skewed.
  2. For the Nested Loops operation, only 100 rows are assumed to be returned by the outer side of the join. Given the high number of rows actually being returned by the MSTVF, you are likely better off with a different join algorithm altogether.
  3. For the Hash Match operation, notice the small warning symbol, which in this case is indicating a spill to disk.

Row flow vs. estimated rows

Contrast the prior plan with the actual plan generated with interleaved execution enabled:

Interleaved plan

  1. Notice that the MSTVF table scan now reflects an accurate cardinality estimate. Also notice the re-ordering of this table scan and the other operations.
  2. And regarding join algorithms, we have switched from a Nested Loop operation to a Hash Match operation instead, which is more optimal given the large number of rows involved.
  3. Also notice that we no longer have spill-warnings, as we're granting more memory based on the true row count flowing from the MSTVF table scan.

Interleaved execution eligible statements

MSTVF referencing statements in interleaved execution must currently be read-only and not part of a data modification operation. Also, MSTVFs are not eligible for interleaved execution if they do not use runtime constants.

Interleaved execution benefits

In general, the higher the skew between the estimated vs. actual number of rows, coupled with the number of downstream plan operations, the greater the performance impact. In general, interleaved execution benefits queries where:

  1. There is a large skew between the estimated vs. actual number of rows for the intermediate result set (in this case, the MSTVF).
  2. And the overall query is sensitive to a change in the size of the intermediate result. This typically happens when there is a complex tree above that subtree in the query plan. A simple SELECT * from an MSTVF will not benefit from interleaved execution.

Interleaved execution overhead

The overhead should be minimal-to-none. MSTVFs were already being materialized prior to the introduction of interleaved execution, however the difference is that now we're now allowing deferred optimization and are then leveraging the cardinality estimate of the materialized row set. As with any plan affecting changes, some plans could change such that with better cardinality for the subtree we get a worse plan for the query overall. Mitigation can include reverting the compatibility level or using Query Store to force the non-regressed version of the plan.

Interleaved execution and consecutive executions

Once an interleaved execution plan is cached, the plan with the revised estimates on the first execution is used for consecutive executions without re-instantiating interleaved execution.

Tracking interleaved execution activity

You can see usage attributes in the actual query execution plan:

Execution Plan attribute Description
ContainsInterleavedExecutionCandidates Applies to the QueryPlan node. When true, means the plan contains interleaved execution candidates.
IsInterleavedExecuted Attribute of the RuntimeInformation element under the RelOp for the TVF node. When true, means the operation was materialized as part of an interleaved execution operation.

You can also track interleaved execution occurrences via the following xEvents:

xEvent Description
interleaved_exec_status This event fires when interleaved execution is occurring.
interleaved_exec_stats_update This event describes the cardinality estimates updated by interleaved execution.
Interleaved_exec_disabled_reason This event fires when a query with a possible candidate for interleaved execution does not actually get interleaved execution.

A query must be executed in order to allow interleaved execution to revise MSTVF cardinality estimates. However, the estimated execution plan still shows when there are interleaved execution candidates via the ContainsInterleavedExecutionCandidates showplan attribute.

Interleaved execution caching

If a plan is cleared or evicted from cache, upon query execution there is a fresh compilation that uses interleaved execution. A statement using OPTION (RECOMPILE) will create a new plan using interleaved execution and not cache it.

Interleaved execution and query store interoperability

Plans using interleaved execution can be forced. The plan is the version that has corrected cardinality estimates based on initial execution.

Disabling interleaved execution without changing the compatibility level

Interleaved execution can be disabled at the database or statement scope while still maintaining database compatibility level 140 and higher. To disable interleaved execution for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET DISABLE_INTERLEAVED_EXECUTION_TVF = ON;

When enabled, this setting will appear as enabled in sys.database_scoped_configurations. To re-enable interleaved execution for all query executions originating from the database, execute the following within the context of the applicable database:

ALTER DATABASE SCOPED CONFIGURATION SET DISABLE_INTERLEAVED_EXECUTION_TVF = OFF;

You can also disable interleaved execution for a specific query by designating DISABLE_INTERLEAVED_EXECUTION_TVF as a USE HINT query hint. For example:

SELECT [fo].[Order Key], [fo].[Quantity], [foo].[OutlierEventQuantity]
FROM [Fact].[Order] AS [fo]
INNER JOIN [Fact].[WhatIfOutlierEventQuantity]('Mild Recession',
                            '1-01-2013',
                            '10-15-2014') AS [foo] ON [fo].[Order Key] = [foo].[Order Key]
                            AND [fo].[City Key] = [foo].[City Key]
                            AND [fo].[Customer Key] = [foo].[Customer Key]
                            AND [fo].[Stock Item Key] = [foo].[Stock Item Key]
                            AND [fo].[Order Date Key] = [foo].[Order Date Key]
                            AND [fo].[Picked Date Key] = [foo].[Picked Date Key]
                            AND [fo].[Salesperson Key] = [foo].[Salesperson Key]
                            AND [fo].[Picker Key] = [foo].[Picker Key]
OPTION (USE HINT('DISABLE_INTERLEAVED_EXECUTION_TVF'));

A USE HINT query hint takes precedence over a database scoped configuration or trace flag setting.

Table variable deferred compilation

Note

Table variable deferred compilation is a public preview feature.

Table variable deferred compilation improves plan quality and overall performance for queries that reference table variables. During optimization and initial compilation, this feature propagates cardinality estimates that are based on actual table variable row counts. This accurate row count information optimizes downstream plan operations.

Table variable deferred compilation defers compilation of a statement that references a table variable until the first actual run of the statement. This deferred compilation behavior is the same as that of temporary tables. This change results in the use of actual cardinality instead of the original one-row guess.

You can enable the public preview of table variable deferred compilation in Azure SQL Database. To do that, enable compatibility level 150 for the database you're connected to when you run the query.

For more information, see Table variable deferred compilation.

Scalar UDF inlining

Note

Scalar user-defined function (UDF) inlining is a public preview feature.

Scalar UDF inlining automatically transforms scalar UDFs into relational expressions. It embeds them in the calling SQL query. This transformation improves the performance of workloads that take advantage of scalar UDFs. Scalar UDF inlining facilitates cost-based optimization of operations inside UDFs. The results are efficient, set-oriented, and parallel instead of inefficient, iterative, serial execution plans. This feature is enabled by default under database compatibility level 150.

For more information, see Scalar UDF inlining.

Approximate query processing

Note

APPROX_COUNT_DISTINCT is a public preview feature.

Approximate query processing is a new feature family. It aggregates across large datasets where responsiveness is more critical than absolute precision. An example is calculating a COUNT(DISTINCT()) across 10 billion rows, for display on a dashboard. In this case, absolute precision isn't important, but responsiveness is critical. The new APPROX_COUNT_DISTINCT aggregate function returns the approximate number of unique non-null values in a group.

For more information, see APPROX_COUNT_DISTINCT (Transact-SQL).

Batch mode on rowstore

Note

Batch mode on rowstore is a public preview feature.

Batch mode on rowstore enables batch mode execution for analytic workloads without requiring columnstore indexes. This feature supports batch mode execution and bitmap filters for on-disk heaps and B-tree indexes. Batch mode on rowstore enables support for all existing batch mode-enabled operators.

Background

SQL Server 2012 (11.x) introduced a new feature to accelerate analytical workloads: columnstore indexes. We expanded the use cases and improved the performance of columnstore indexes in each subsequent release. Until now, we surfaced and documented all these capabilities as a single feature. You create columnstore indexes on your tables. And your analytical workload goes faster. However, there are two related but distinct sets of technologies:

  • With columnstore indexes, analytical queries access only the data in the columns they need. Page compression in the columnstore format is also more effective than compression in traditional rowstore indexes.
  • With batch mode processing, query operators process data more efficiently. They work on a batch of rows instead of one row at a time. A number of other scalability improvements are tied to batch mode processing. For more information on batch mode, see Execution modes.

The two sets of features work together to improve input/output (I/O) and CPU use:

  • By using columnstore indexes, more of your data fits in memory. That reduces the need for I/O.
  • Batch mode processing uses CPU more efficiently.

The two technologies take advantage of each other whenever possible. For example, batch mode aggregates can be evaluated as part of a columnstore index scan. We also process columnstore data that's compressed by using run-length encoding much more efficiently with batch mode joins and batch mode aggregates.

The two features are independently usable:

  • You get row mode plans that use columnstore indexes.
  • You get batch mode plans that use only rowstore indexes.

You usually get the best results when you use the two features together. So until now, the SQL Server query optimizer considered batch mode processing only for queries that involve at least one table with a columnstore index.

Columnstore indexes aren't a good option for some applications. An application might use some other feature that isn't supported with columnstore indexes. For example, in-place modifications aren't compatible with columnstore compression. So triggers aren't supported on tables with clustered columnstore indexes. More important, columnstore indexes add overhead for DELETE and UPDATE statements.

For some hybrid transactional-analytical workloads, the overhead on a workload's transactional aspects outweighs the benefits of columnstore indexes. Such scenarios can improve CPU use from batch mode processing alone. That's why the batch mode on rowstore feature considers batch mode for all queries. It doesn't matter which indexes are involved.

Workloads that might benefit from batch mode on rowstore

The following workloads might benefit from batch mode on rowstore:

  • A significant part of the workload consists of analytical queries. Usually, these queries have operators like joins or aggregates that process hundreds of thousands of rows or more.
  • The workload is CPU bound. If the bottleneck is I/O, we still recommend that you consider a columnstore index, if possible.
  • Creating a columnstore index adds too much overhead to the transactional part of your workload. Or, creating a columnstore index isn't feasible because your application depends on a feature that's not yet supported with columnstore indexes.

Note

Batch mode on rowstore helps only by reducing CPU consumption. If your bottleneck is I/O related, and data isn't already cached ("cold" cache), batch mode on rowstore won't improve elapsed time. Similarly, if there isn't enough memory on the machine to cache all the data, a performance improvement is unlikely.

What changes with batch mode on rowstore?

Other than moving to compatibility level 150, you don't have to change anything on your side to enable batch mode on rowstore for candidate workloads.

Even if a query doesn't involve any table with a columnstore index, the query processor now uses heuristics to decide whether to consider batch mode. The heuristics consist of these checks:

  1. An initial check of table sizes, operators used, and estimated cardinalities in the input query.
  2. Additional checkpoints, as the optimizer discovers new, cheaper plans for the query. If these alternative plans don't make significant use of batch mode, the optimizer stops exploring batch mode alternatives.

If batch mode on rowstore is used, you see the actual run mode as batch mode in the query plan. The scan operator uses batch mode for on-disk heaps and B-tree indexes. This batch mode scan can evaluate batch mode bitmap filters. You might also see other batch mode operators in the plan. Examples are hash joins, hash-based aggregates, sorts, window aggregates, filters, concatenation, and compute scalar operators.

Remarks

  • Query plans don't always use batch mode. The query optimizer might decide that batch mode isn't beneficial for the query.
  • The query optimizer's search space is changing. So if you get a row mode plan, it might not be the same as the plan you get in a lower compatibility level. And if you get a batch mode plan, it might not be the same as the plan you get with a columnstore index.
  • Plans might also change for queries that mix columnstore and rowstore indexes because of the new batch mode rowstore scan.
  • There are current limitations for the new batch mode on rowstore scan:
    • It won't kick in for in-memory OLTP tables or for any index other than on-disk heaps and B-trees.
    • It also won't kick in if a large object (LOB) column is fetched or filtered. This limitation includes sparse column sets and XML columns.
  • There are queries that batch mode isn't used for even with columnstore indexes. Examples are queries that involve cursors. These same exclusions also extend to batch mode on rowstore.

Configure batch mode on rowstore

The BATCH_MODE_ON_ROWSTORE database scoped configuration is on by default. It disables batch mode on rowstore without requiring a change in database compatibility level:

-- Disabling batch mode on rowstore
ALTER DATABASE SCOPED CONFIGURATION SET BATCH_MODE_ON_ROWSTORE = OFF;

-- Enabling batch mode on rowstore
ALTER DATABASE SCOPED CONFIGURATION SET BATCH_MODE_ON_ROWSTORE = ON;

You can disable batch mode on rowstore via database scoped configuration. But you can still override the setting at the query level by using the ALLOW_BATCH_MODE query hint. The following example enables batch mode on rowstore even with the feature disabled via database scoped configuration:

SELECT [Tax Rate], [Lineage Key], [Salesperson Key], SUM(Quantity) AS SUM_QTY, SUM([Unit Price]) AS SUM_BASE_PRICE, COUNT(*) AS COUNT_ORDER
FROM Fact.OrderHistoryExtended
WHERE [Order Date Key]<=DATEADD(dd, -73, '2015-11-13')
GROUP BY [Tax Rate], [Lineage Key], [Salesperson Key]
ORDER BY [Tax Rate], [Lineage Key], [Salesperson Key]
OPTION(RECOMPILE, USE HINT('ALLOW_BATCH_MODE'));

You can also disable batch mode on rowstore for a specific query by using the DISALLOW_BATCH_MODE query hint. See the following example:

SELECT [Tax Rate], [Lineage Key], [Salesperson Key], SUM(Quantity) AS SUM_QTY, SUM([Unit Price]) AS SUM_BASE_PRICE, COUNT(*) AS COUNT_ORDER
FROM Fact.OrderHistoryExtended
WHERE [Order Date Key]<=DATEADD(dd, -73, '2015-11-13')
GROUP BY [Tax Rate], [Lineage Key], [Salesperson Key]
ORDER BY [Tax Rate], [Lineage Key], [Salesperson Key]
OPTION(RECOMPILE, USE HINT('DISALLOW_BATCH_MODE'));

See also

Performance Center for SQL Server Database Engine and Azure SQL Database
Query processing architecture guide
Showplan logical and physical operators reference
Joins
Demonstrating adaptive query processing
Demonstrating intelligent QP