Skip to content
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

util/ranger: add missing Selection for range scan from like on PAD SPACE column | tidb-test=pr/2251 (#48845) #48882

Closed
wants to merge 1 commit into from
Closed

util/ranger: add missing Selection for range scan from like on PAD SPACE column | tidb-test=pr/2251 (#48845) #48882

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
18 changes: 9 additions & 9 deletions pkg/planner/core/casetest/physicalplantest/testdata/plan_suite_out.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2171,11 +2171,11 @@
},
{
"SQL": "select a from t where c_str like ''",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"\",\"\"]])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"\",\"\"]]->Sel([like(test.t.c_str, , 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc'",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc\",\"abc\"]])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc\",\"abc\"]]->Sel([like(test.t.c_str, abc, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str not like 'abc'",
Expand All @@ -2191,7 +2191,7 @@
},
{
"SQL": "select a from t where c_str like 'abc%'",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc\",\"abd\")])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc\",\"abd\")]->Sel([like(test.t.c_str, abc%, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc_'",
Expand All @@ -2203,31 +2203,31 @@
},
{
"SQL": "select a from t where c_str like 'abc\\_' escape ''",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc_\"]])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc_\"]]->Sel([like(test.t.c_str, abc\\_, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc\\_'",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc_\"]])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc_\"]]->Sel([like(test.t.c_str, abc\\_, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc\\\\_'",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc_\"]])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc_\"]]->Sel([like(test.t.c_str, abc\\_, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc\\_%'",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc`\")])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc`\")]->Sel([like(test.t.c_str, abc\\_%, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc=_%' escape '='",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc`\")])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"abc_\",\"abc`\")]->Sel([like(test.t.c_str, abc=_%, 61)]))->Projection"
},
{
"SQL": "select a from t where c_str like 'abc\\__'",
"Best": "IndexReader(Index(t.c_d_e_str)[(\"abc_\",\"abc`\")]->Sel([like(test.t.c_str, abc\\__, 92)]))->Projection"
},
{
"SQL": "select a from t where c_str like 123",
"Best": "IndexReader(Index(t.c_d_e_str)[[\"123\",\"123\"]])->Projection"
"Best": "IndexReader(Index(t.c_d_e_str)[[\"123\",\"123\"]]->Sel([like(test.t.c_str, 123, 92)]))->Projection"
},
{
"SQL": "select a from t where c = 1.9 and d > 3",
Expand Down
16 changes: 14 additions & 2 deletions pkg/util/ranger/checker.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@ package ranger
import (
"github.com/pingcap/tidb/pkg/expression"
"github.com/pingcap/tidb/pkg/parser/ast"
"github.com/pingcap/tidb/pkg/parser/charset"
"github.com/pingcap/tidb/pkg/parser/mysql"
"github.com/pingcap/tidb/pkg/types"
"github.com/pingcap/tidb/pkg/util/collate"
Expand Down Expand Up @@ -166,11 +167,22 @@ func (c *conditionChecker) checkLikeFunc(scalar *expression.ScalarFunction) (isA
if err != nil {
return false, true
}
likeFuncReserve := !c.isFullLengthColumn()

// Different from `=`, trailing spaces are always significant, and can't be ignored in `like`.
// In tidb's implementation, for PAD SPACE collations, the trailing spaces are removed in the index key. So we are
// unable to distinguish 'xxx' from 'xxx ' by a single index range scan, and we may read more data than needed by
// the `like` function. Therefore, a Selection is needed to filter the data.
// Since all collations, except for binary, implemented in tidb are PAD SPACE collations for now, we use a simple
// collation != binary check here.
if collation != charset.CollationBin {
likeFuncReserve = true
}

if len(patternStr) == 0 {
return true, !c.isFullLengthColumn()
return true, likeFuncReserve
}
escape := byte(scalar.GetArgs()[2].(*expression.Constant).Value.GetInt64())
likeFuncReserve := !c.isFullLengthColumn()
for i := 0; i < len(patternStr); i++ {
if patternStr[i] == escape {
i++
Expand Down
12 changes: 6 additions & 6 deletions pkg/util/ranger/ranger_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1098,7 +1098,7 @@ create table t(
indexPos: 0,
exprStr: `a LIKE 'abc%'`,
accessConds: `[like(test.t.a, abc%, 92)]`,
filterConds: "[]",
filterConds: "[like(test.t.a, abc%, 92)]",
resultStr: "[[\"abc\",\"abd\")]",
},
{
Expand All @@ -1112,14 +1112,14 @@ create table t(
indexPos: 0,
exprStr: "a LIKE 'abc'",
accessConds: "[like(test.t.a, abc, 92)]",
filterConds: "[]",
filterConds: "[like(test.t.a, abc, 92)]",
resultStr: "[[\"abc\",\"abc\"]]",
},
{
indexPos: 0,
exprStr: `a LIKE "ab\_c"`,
accessConds: "[like(test.t.a, ab\\_c, 92)]",
filterConds: "[]",
filterConds: "[like(test.t.a, ab\\_c, 92)]",
resultStr: "[[\"ab_c\",\"ab_c\"]]",
},
{
Expand All @@ -1133,21 +1133,21 @@ create table t(
indexPos: 0,
exprStr: `a LIKE '\%a'`,
accessConds: "[like(test.t.a, \\%a, 92)]",
filterConds: "[]",
filterConds: "[like(test.t.a, \\%a, 92)]",
resultStr: `[["%a","%a"]]`,
},
{
indexPos: 0,
exprStr: `a LIKE "\\"`,
accessConds: "[like(test.t.a, \\, 92)]",
filterConds: "[]",
filterConds: "[like(test.t.a, \\, 92)]",
resultStr: "[[\"\\\\\",\"\\\\\"]]",
},
{
indexPos: 0,
exprStr: `a LIKE "\\\\a%"`,
accessConds: `[like(test.t.a, \\a%, 92)]`,
filterConds: "[]",
filterConds: "[like(test.t.a, \\\\a%, 92)]",
resultStr: "[[\"\\\\a\",\"\\\\b\")]",
},
{
Expand Down
3 changes: 2 additions & 1 deletion tests/integrationtest/r/explain_generate_column_substitute.result
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -415,7 +415,8 @@ id estRows task access object operator info
StreamAgg 1.00 root funcs:count(Column#6)->Column#4
└─IndexReader 1.00 root index:StreamAgg
└─StreamAgg 1.00 cop[tikv] funcs:count(1)->Column#6
└─IndexRangeScan 250.00 cop[tikv] table:tbl1, index:expression_index(md5(`s`)) range:["02e74f10e0327ad868d138f2b4fdd6f","02e74f10e0327ad868d138f2b4fdd6g"), keep order:false, stats:pseudo
└─Selection 250.00 cop[tikv] like(md5(cast(explain_generate_column_substitute.tbl1.s, var_string(20))), "02e74f10e0327ad868d138f2b4fdd6f%", 92)
└─IndexRangeScan 250.00 cop[tikv] table:tbl1, index:expression_index(md5(`s`)) range:["02e74f10e0327ad868d138f2b4fdd6f","02e74f10e0327ad868d138f2b4fdd6g"), keep order:false, stats:pseudo
select count(*) from tbl1 use index() where md5(s) like '02e74f10e0327ad868d138f2b4fdd6f%';
count(*)
64
Expand Down
Loading