Skip to content

planner: fix LogicalProjection.DeriveStats allocate too many memories (#63829) | tidb-test=920d9bf1b1137cda1272bdd59ae527aee8067944 #64006

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
ti-chi-bot merged 3 commits into from
Oct 17, 2025
Merged

planner: fix LogicalProjection.DeriveStats allocate too many memories (#63829) | tidb-test=920d9bf1b1137cda1272bdd59ae527aee8067944 #64006

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
564a0ca
This is an automated cherry-pick of #63829
hawkingrei Oct 9, 2025
8362afb
update
hawkingrei Oct 17, 2025
4cc0656
update
hawkingrei Oct 17, 2025
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
32 changes: 32 additions & 0 deletions expression/util.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,10 +35,12 @@ import (
"github.com/pingcap/tidb/types"
driver "github.com/pingcap/tidb/types/parser_driver"
"github.com/pingcap/tidb/util/chunk"
"github.com/pingcap/tidb/util/cmp"
"github.com/pingcap/tidb/util/collate"
"github.com/pingcap/tidb/util/logutil"
"github.com/pingcap/tidb/util/sqlexec"
"go.uber.org/zap"
"golang.org/x/exp/slices"
)

// cowExprRef is a copy-on-write slice ref util using in `ColumnSubstitute`
Expand Down Expand Up @@ -122,6 +124,36 @@ func ExtractColumns(expr Expression) []*Column {
return extractColumns(result, expr, nil)
}

// ExtractAllColumnsFromExpressionsInUsedSlices is the same as ExtractColumns. but it can reuse the memory.
func ExtractAllColumnsFromExpressionsInUsedSlices(reuse []*Column, exprs ...Expression) []*Column {
if len(exprs) == 0 {
return nil
}
reuse = reuse[:0]
for _, expr := range exprs {
Copy link
Contributor

@AilinKid AilinKid Oct 17, 2025

Choose a reason for hiding this comment

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

cols = cols[:0]

Copy link
Contributor

@AilinKid AilinKid Oct 17, 2025 •
edited
Loading

Choose a reason for hiding this comment

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

how about move to here, after all, your func name says it will reuse...

Copy link
Member

@hawkingrei hawkingrei Oct 17, 2025

Choose a reason for hiding this comment

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

Updated

reuse = extractColumnsSlices(reuse, expr)
}
slices.SortFunc(reuse, func(a, b *Column) int {
return cmp.Compare(a.UniqueID, b.UniqueID)
})
reuse = slices.CompactFunc(reuse, func(a, b *Column) bool {
return a.UniqueID == b.UniqueID
})
return reuse
}

func extractColumnsSlices(result []*Column, expr Expression) []*Column {
switch v := expr.(type) {
case *Column:
result = append(result, v)
case *ScalarFunction:
for _, arg := range v.GetArgs() {
result = extractColumnsSlices(result, arg)
}
}
return result
}

// ExtractCorColumns extracts correlated column from given expression.
func ExtractCorColumns(expr Expression) (cols []*CorrelatedColumn) {
switch v := expr.(type) {
Expand Down
3 changes: 2 additions & 1 deletion planner/core/stats.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1127,8 +1127,9 @@ func (p *LogicalProjection) DeriveStats(childStats []*property.StatsInfo, selfSc
RowCount: childProfile.RowCount,
ColNDVs: make(map[int64]float64, len(p.Exprs)),
}
cols := make([]*expression.Column, 0, 8)
for i, expr := range p.Exprs {
cols := expression.ExtractColumns(expr)
cols = expression.ExtractAllColumnsFromExpressionsInUsedSlices(cols, expr)
p.stats.ColNDVs[selfSchema.Columns[i].UniqueID], _ = getColsNDVWithMatchedLen(cols, childSchema[0], childProfile)
}
p.stats.GroupNDVs = p.getGroupNDVs(colGroups, childProfile, selfSchema)
Expand Down