Skip to content

Optimize MIN/MAX over DISTINCT #34699

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Sep 17, 2024
Merged

Optimize MIN/MAX over DISTINCT #34699

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
158fdc8
Optimize away `DISTINCT` in `MIN`/`MAX`
ranma42 Aug 7, 2024
8413964
Optimize MIN/MAX on DISTINCT `EnumerableExpression`s
ranma42 Aug 7, 2024
6820eb6
Update baselines
ranma42 Aug 7, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 4 additions & 3 deletions src/EFCore.Relational/Query/EnumerableExpression.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,11 +68,12 @@ public virtual EnumerableExpression ApplySelector(Expression expression)
=> new(expression, IsDistinct, Predicate, Orderings);

/// <summary>
/// Applies DISTINCT operator to the selector of the <see cref="EnumerableExpression" />.
/// Sets whether the DISTINCT operator should be applied to the selector
/// of the <see cref="EnumerableExpression" />.
/// </summary>
/// <returns>The new expression with specified component updated.</returns>
public virtual EnumerableExpression ApplyDistinct()
=> new(Selector, distinct: true, Predicate, Orderings);
public virtual EnumerableExpression SetDistinct(bool value)
=> new(Selector, distinct: value, Predicate, Orderings);

/// <summary>
/// Applies filter predicate to the <see cref="EnumerableExpression" />.
Expand Down
6 changes: 6 additions & 0 deletions src/EFCore.Relational/Query/RelationalQueryableMethodTranslatingExpressionVisitor.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -954,6 +954,9 @@ private SqlExpression CreateJoinPredicate(Expression outerKey, Expression innerK
/// <inheritdoc />
protected override ShapedQueryExpression? TranslateMax(ShapedQueryExpression source, LambdaExpression? selector, Type resultType)
{
var selectExpression = (SelectExpression)source.QueryExpression;
selectExpression.IsDistinct = false;

// For Max() over an inline array, translate to GREATEST() if possible; otherwise use the default translation of aggregate SQL
// MAX().
// Note that some providers propagate NULL arguments (SQLite, MySQL), while others only return NULL if all arguments evaluate to
Expand All @@ -974,6 +977,9 @@ private SqlExpression CreateJoinPredicate(Expression outerKey, Expression innerK
/// <inheritdoc />
protected override ShapedQueryExpression? TranslateMin(ShapedQueryExpression source, LambdaExpression? selector, Type resultType)
{
var selectExpression = (SelectExpression)source.QueryExpression;
selectExpression.IsDistinct = false;

// See comments above in TranslateMax()
if (TryExtractBareInlineCollectionValues(source, out var values)
&& _sqlTranslator.GenerateLeast(values, resultType.UnwrapNullableType()) is SqlFunctionExpression leastExpression
Expand Down
17 changes: 12 additions & 5 deletions src/EFCore.Relational/Query/RelationalSqlTranslatingExpressionVisitor.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1362,12 +1362,15 @@ protected virtual bool TryTranslateAggregateMethodCall(
{
switch (genericMethod.Name)
{
case nameof(Queryable.Average)
when QueryableMethods.IsAverageWithoutSelector(genericMethod):
case nameof(Queryable.Max)
when genericMethod == QueryableMethods.MaxWithoutSelector:
case nameof(Queryable.Min)
when genericMethod == QueryableMethods.MinWithoutSelector:
enumerableExpression = enumerableExpression.SetDistinct(false);
break;

case nameof(Queryable.Average)
when QueryableMethods.IsAverageWithoutSelector(genericMethod):
case nameof(Queryable.Sum)
when QueryableMethods.IsSumWithoutSelector(genericMethod):
case nameof(Queryable.Count)
Expand All @@ -1376,12 +1379,16 @@ when QueryableMethods.IsSumWithoutSelector(genericMethod):
when genericMethod == QueryableMethods.LongCountWithoutPredicate:
break;

case nameof(Queryable.Average)
when QueryableMethods.IsAverageWithSelector(genericMethod):
case nameof(Queryable.Max)
when genericMethod == QueryableMethods.MaxWithSelector:
case nameof(Queryable.Min)
when genericMethod == QueryableMethods.MinWithSelector:
enumerableExpression = enumerableExpression.SetDistinct(false);
enumerableExpression = ProcessSelector(enumerableExpression, arguments[1].UnwrapLambdaFromQuote());
break;

case nameof(Queryable.Average)
when QueryableMethods.IsAverageWithSelector(genericMethod):
case nameof(Queryable.Sum)
when QueryableMethods.IsSumWithSelector(genericMethod):
enumerableExpression = ProcessSelector(enumerableExpression, arguments[1].UnwrapLambdaFromQuote());
Expand Down Expand Up @@ -1478,7 +1485,7 @@ private bool TryTranslateAsEnumerableExpression(

if (!enumerableSource.IsDistinct)
{
enumerableExpression = enumerableSource.ApplyDistinct();
enumerableExpression = enumerableSource.SetDistinct(true);
return true;
}

Expand Down
8 changes: 4 additions & 4 deletions test/EFCore.SqlServer.FunctionalTests/Query/NorthwindGroupByQuerySqlServerTest.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2475,7 +2475,7 @@ public override async Task GroupBy_Select_Distinct_aggregate(bool async)

AssertSql(
"""
SELECT [o].[CustomerID] AS [Key], AVG(DISTINCT (CAST([o].[OrderID] AS float))) AS [Average], COUNT(DISTINCT ([o].[EmployeeID])) AS [Count], COUNT_BIG(DISTINCT ([o].[EmployeeID])) AS [LongCount], MAX(DISTINCT ([o].[OrderDate])) AS [Max], MIN(DISTINCT ([o].[OrderDate])) AS [Min], COALESCE(SUM(DISTINCT ([o].[OrderID])), 0) AS [Sum]
SELECT [o].[CustomerID] AS [Key], AVG(DISTINCT (CAST([o].[OrderID] AS float))) AS [Average], COUNT(DISTINCT ([o].[EmployeeID])) AS [Count], COUNT_BIG(DISTINCT ([o].[EmployeeID])) AS [LongCount], MAX([o].[OrderDate]) AS [Max], MIN([o].[OrderDate]) AS [Min], COALESCE(SUM(DISTINCT ([o].[OrderID])), 0) AS [Sum]
FROM [Orders] AS [o]
GROUP BY [o].[CustomerID]
""");
Expand All @@ -2487,7 +2487,7 @@ public override async Task GroupBy_group_Distinct_Select_Distinct_aggregate(bool

AssertSql(
"""
SELECT [o].[CustomerID] AS [Key], MAX(DISTINCT ([o].[OrderDate])) AS [Max]
SELECT [o].[CustomerID] AS [Key], MAX([o].[OrderDate]) AS [Max]
FROM [Orders] AS [o]
GROUP BY [o].[CustomerID]
""");
Expand All @@ -2499,9 +2499,9 @@ public override async Task GroupBy_group_Where_Select_Distinct_aggregate(bool as

AssertSql(
"""
SELECT [o].[CustomerID] AS [Key], MAX(DISTINCT (CASE
SELECT [o].[CustomerID] AS [Key], MAX(CASE
WHEN [o].[OrderDate] IS NOT NULL THEN [o].[OrderDate]
END)) AS [Max]
END) AS [Max]
FROM [Orders] AS [o]
GROUP BY [o].[CustomerID]
""");
Expand Down
14 changes: 4 additions & 10 deletions test/EFCore.SqlServer.FunctionalTests/Query/NorthwindMiscellaneousQuerySqlServerTest.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4744,11 +4744,8 @@ public override async Task Select_distinct_max(bool async)

AssertSql(
"""
SELECT MAX([o0].[OrderID])
FROM (
SELECT DISTINCT [o].[OrderID]
FROM [Orders] AS [o]
) AS [o0]
SELECT MAX([o].[OrderID])
Copy link
Contributor Author

@ranma42 ranma42 Sep 17, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

these tests seem to be aimed at checking exactly this behavior; should I add something similar or are they adequate?

Copy link
Member

@roji roji Sep 17, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

They look good to me...

FROM [Orders] AS [o]
""");
}

Expand All @@ -4758,11 +4755,8 @@ public override async Task Select_distinct_min(bool async)

AssertSql(
"""
SELECT MIN([o0].[OrderID])
FROM (
SELECT DISTINCT [o].[OrderID]
FROM [Orders] AS [o]
) AS [o0]
SELECT MIN([o].[OrderID])
FROM [Orders] AS [o]
""");
}

Expand Down