expression:implement vectorized evaluation for builtinConvertSig (pin…

…gcap#13537)
qw4990 · Nov 27, 2019 · a85a9f9 · a85a9f9
1 parent 01a9e13
commit a85a9f9
Show file tree

Hide file tree

Showing 2 changed files with 56 additions and 3 deletions.
diff --git a/expression/builtin_string_vec.go b/expression/builtin_string_vec.go
@@ -24,9 +24,11 @@ import (
 
 	"github.com/pingcap/errors"
 	"github.com/pingcap/parser/ast"
+	"github.com/pingcap/parser/charset"
 	"github.com/pingcap/parser/mysql"
 	"github.com/pingcap/tidb/types"
 	"github.com/pingcap/tidb/util/chunk"
+	"golang.org/x/text/transform"
 )
 
 func (b *builtinLowerSig) vecEvalString(input *chunk.Chunk, result *chunk.Column) error {
@@ -599,11 +601,41 @@ func (b *builtinConcatWSSig) vecEvalString(input *chunk.Chunk, result *chunk.Col
 }
 
 func (b *builtinConvertSig) vectorized() bool {
-	return false
+	return true
 }
 
 func (b *builtinConvertSig) vecEvalString(input *chunk.Chunk, result *chunk.Column) error {
-	return errors.Errorf("not implemented")
+	n := input.NumRows()
+	expr, err := b.bufAllocator.get(types.ETString, n)
+	if err != nil {
+		return err
+	}
+	defer b.bufAllocator.put(expr)
+	if err := b.args[0].VecEvalString(b.ctx, input, expr); err != nil {
+		return err
+	}
+	// Since charset is already validated and set from getFunction(), there's no
+	// need to get charset from args again.
+	encoding, _ := charset.Lookup(b.tp.Charset)
+	// However, if `b.tp.Charset` is abnormally set to a wrong charset, we still
+	// return with error.
+	if encoding == nil {
+		return errUnknownCharacterSet.GenWithStackByArgs(b.tp.Charset)
+	}
+	result.ReserveString(n)
+	for i := 0; i < n; i++ {
+		if expr.IsNull(i) {
+			result.AppendNull()
+			continue
+		}
+		exprI := expr.GetString(i)
+		target, _, err := transform.String(encoding.NewDecoder(), exprI)
+		if err != nil {
+			return err
+		}
+		result.AppendString(target)
+	}
+	return nil
 }
 
 func (b *builtinSubstringIndexSig) vectorized() bool {

diff --git a/expression/builtin_string_vec_test.go b/expression/builtin_string_vec_test.go
@@ -58,7 +58,28 @@ var vecBuiltinStringCases = map[string][]vecExprBenchCase{
 		{retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETString, types.ETString}},
 	},
 	ast.ConcatWS: {},
-	ast.Convert:  {},
+	ast.Convert: {
+		{
+			retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETString},
+			constants: []*Constant{nil, {Value: types.NewDatum("utf8"), RetType: types.NewFieldType(mysql.TypeString)}},
+		},
+		{
+			retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETString},
+			constants: []*Constant{nil, {Value: types.NewDatum("binary"), RetType: types.NewFieldType(mysql.TypeString)}},
+		},
+		{
+			retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETString},
+			constants: []*Constant{nil, {Value: types.NewDatum("utf8mb4"), RetType: types.NewFieldType(mysql.TypeString)}},
+		},
+		{
+			retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETString},
+			constants: []*Constant{nil, {Value: types.NewDatum("ascii"), RetType: types.NewFieldType(mysql.TypeString)}},
+		},
+		{
+			retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETString},
+			constants: []*Constant{nil, {Value: types.NewDatum("latin1"), RetType: types.NewFieldType(mysql.TypeString)}},
+		},
+	},
 	ast.Substring: {
 		{
 			retEvalType:   types.ETString,