expression, parser: add built-in func is_uuid (#30318)

Author: unconsolable

executor/show_test.go

@@ -1394,7 +1394,7 @@ func (s *testSuite5) TestShowBuiltin(c *C) {
 	res := tk.MustQuery("show builtins;")
 	c.Assert(res, NotNil)
 	rows := res.Rows()
-	const builtinFuncNum = 273
+	const builtinFuncNum = 274
 	c.Assert(builtinFuncNum, Equals, len(rows))
 	c.Assert("abs", Equals, rows[0][0].(string))
 	c.Assert("yearweek", Equals, rows[builtinFuncNum-1][0].(string))

expression/builtin.go

@@ -772,6 +772,7 @@ var funcs = map[string]functionClass{
 	ast.IsIPv4Mapped:    &isIPv4MappedFunctionClass{baseFunctionClass{ast.IsIPv4Mapped, 1, 1}},
 	ast.IsIPv6:          &isIPv6FunctionClass{baseFunctionClass{ast.IsIPv6, 1, 1}},
 	ast.IsUsedLock:      &isUsedLockFunctionClass{baseFunctionClass{ast.IsUsedLock, 1, 1}},
+	ast.IsUUID:          &isUUIDFunctionClass{baseFunctionClass{ast.IsUUID, 1, 1}},
 	ast.MasterPosWait:   &masterPosWaitFunctionClass{baseFunctionClass{ast.MasterPosWait, 2, 4}},
 	ast.NameConst:       &nameConstFunctionClass{baseFunctionClass{ast.NameConst, 2, 2}},
 	ast.ReleaseAllLocks: &releaseAllLocksFunctionClass{baseFunctionClass{ast.ReleaseAllLocks, 0, 0}},

expression/builtin_miscellaneous.go

@@ -57,6 +57,7 @@ var (
 	_ functionClass = &vitessHashFunctionClass{}
 	_ functionClass = &uuidToBinFunctionClass{}
 	_ functionClass = &binToUUIDFunctionClass{}
+	_ functionClass = &isUUIDFunctionClass{}
 )
 
 var (
@@ -78,6 +79,7 @@ var (
 	_ builtinFunc = &builtinIsIPv4CompatSig{}
 	_ builtinFunc = &builtinIsIPv4MappedSig{}
 	_ builtinFunc = &builtinIsIPv6Sig{}
+	_ builtinFunc = &builtinIsUUIDSig{}
 	_ builtinFunc = &builtinUUIDSig{}
 	_ builtinFunc = &builtinVitessHashSig{}
 	_ builtinFunc = &builtinUUIDToBinSig{}
@@ -862,6 +864,47 @@ func (c *isUsedLockFunctionClass) getFunction(ctx sessionctx.Context, args []Exp
 	return nil, errFunctionNotExists.GenWithStackByArgs("FUNCTION", "IS_USED_LOCK")
 }
 
+type isUUIDFunctionClass struct {
+	baseFunctionClass
+}
+
+func (c *isUUIDFunctionClass) getFunction(ctx sessionctx.Context, args []Expression) (builtinFunc, error) {
+	if err := c.verifyArgs(args); err != nil {
+		return nil, err
+	}
+	bf, err := newBaseBuiltinFuncWithTp(ctx, c.funcName, args, types.ETInt, types.ETString)
+	if err != nil {
+		return nil, err
+	}
+	bf.tp.Flen = 1
+	sig := &builtinIsUUIDSig{bf}
+	sig.setPbCode(tipb.ScalarFuncSig_IsUUID)
+	return sig, nil
+}
+
+type builtinIsUUIDSig struct {
+	baseBuiltinFunc
+}
+
+func (b *builtinIsUUIDSig) Clone() builtinFunc {
+	newSig := &builtinIsUUIDSig{}
+	newSig.cloneFrom(&b.baseBuiltinFunc)
+	return newSig
+}
+
+// evalInt evals a builtinIsUUIDSig.
+// See https://dev.mysql.com/doc/refman/8.0/en/miscellaneous-functions.html#function_is-uuid
+func (b *builtinIsUUIDSig) evalInt(row chunk.Row) (int64, bool, error) {
+	val, isNull, err := b.args[0].EvalString(b.ctx, row)
+	if err != nil || isNull {
+		return 0, isNull, err
+	}
+	if _, err = uuid.Parse(val); err != nil {
+		return 0, false, nil
+	}
+	return 1, false, nil
+}
+
 type masterPosWaitFunctionClass struct {
 	baseFunctionClass
 }

expression/builtin_miscellaneous_test.go

@@ -97,6 +97,42 @@ func TestIsIPv4(t *testing.T) {
 	trequire.DatumEqual(t, types.NewDatum(0), r)
 }
 
+func TestIsUUID(t *testing.T) {
+	t.Parallel()
+	ctx := createContext(t)
+	tests := []struct {
+		uuid   string
+		expect interface{}
+	}{
+		{"6ccd780c-baba-1026-9564-5b8c656024db", 1},
+		{"6CCD780C-BABA-1026-9564-5B8C656024DB", 1},
+		{"6ccd780cbaba102695645b8c656024db", 1},
+		{"{6ccd780c-baba-1026-9564-5b8c656024db}", 1},
+		{"6ccd780c-baba-1026-9564-5b8c6560", 0},
+		{"6CCD780C-BABA-1026-9564-5B8C656024DQ", 0},
+		// This is a bug in google/uuid#60
+		{"{99a9ad03-5298-11ec-8f5c-00ff90147ac3*", 1},
+		// This is a format google/uuid support, while mysql doesn't
+		{"urn:uuid:99a9ad03-5298-11ec-8f5c-00ff90147ac3", 1},
+	}
+
+	fc := funcs[ast.IsUUID]
+	for _, test := range tests {
+		uuid := types.NewStringDatum(test.uuid)
+		f, err := fc.getFunction(ctx, datumsToConstants([]types.Datum{uuid}))
+		require.NoError(t, err)
+		result, err := evalBuiltinFunc(f, chunk.Row{})
+		require.NoError(t, err)
+		trequire.DatumEqual(t, types.NewDatum(test.expect), result)
+	}
+
+	var argNull types.Datum
+	f, _ := fc.getFunction(ctx, datumsToConstants([]types.Datum{argNull}))
+	r, err := evalBuiltinFunc(f, chunk.Row{})
+	require.NoError(t, err)
+	require.True(t, r.IsNull())
+}
+
 func TestUUID(t *testing.T) {
 	t.Parallel()
 	ctx := createContext(t)

expression/builtin_miscellaneous_vec.go

@@ -154,6 +154,36 @@ func (b *builtinIsIPv6Sig) vecEvalInt(input *chunk.Chunk, result *chunk.Column)
 	return nil
 }
 
+func (b *builtinIsUUIDSig) vectorized() bool {
+	return true
+}
+
+func (b *builtinIsUUIDSig) vecEvalInt(input *chunk.Chunk, result *chunk.Column) error {
+	n := input.NumRows()
+	buf, err := b.bufAllocator.get()
+	if err != nil {
+		return err
+	}
+	defer b.bufAllocator.put(buf)
+	if err := b.args[0].VecEvalString(b.ctx, input, buf); err != nil {
+		return err
+	}
+	result.ResizeInt64(n, false)
+	i64s := result.Int64s()
+	result.MergeNulls(buf)
+	for i := 0; i < n; i++ {
+		if result.IsNull(i) {
+			continue
+		}
+		if _, err = uuid.Parse(buf.GetString(i)); err != nil {
+			i64s[i] = 0
+		} else {
+			i64s[i] = 1
+		}
+	}
+	return nil
+}
+
 func (b *builtinNameConstStringSig) vectorized() bool {
 	return true
 }

expression/builtin_miscellaneous_vec_test.go

@@ -108,6 +108,9 @@ var vecBuiltinMiscellaneousCases = map[string][]vecExprBenchCase{
 		{retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString}, geners: []dataGenerator{&uuidBinGener{newDefaultRandGen()}}},
 		{retEvalType: types.ETString, childrenTypes: []types.EvalType{types.ETString, types.ETInt}, geners: []dataGenerator{&uuidBinGener{newDefaultRandGen()}}},
 	},
+	ast.IsUUID: {
+		{retEvalType: types.ETInt, childrenTypes: []types.EvalType{types.ETString}, geners: []dataGenerator{&uuidStrGener{newDefaultRandGen()}}},
+	},
 }
 
 func TestVectorizedBuiltinMiscellaneousEvalOneVec(t *testing.T) {

go.mod

@@ -53,7 +53,7 @@ require (
 	github.com/pingcap/sysutil v0.0.0-20210730114356-fcd8a63f68c5
 	github.com/pingcap/tidb-tools v5.2.2-0.20211019062242-37a8bef2fa17+incompatible
 	github.com/pingcap/tidb/parser v0.0.0-20211011031125-9b13dc409c5e
-	github.com/pingcap/tipb v0.0.0-20211116093845-e9b045a0bdf8
+	github.com/pingcap/tipb v0.0.0-20211201080053-bd104bb270ba
 	github.com/prometheus/client_golang v1.5.1
 	github.com/prometheus/client_model v0.2.0
 	github.com/prometheus/common v0.9.1

go.sum

@@ -597,8 +597,8 @@ github.com/pingcap/tidb-dashboard v0.0.0-20211008050453-a25c25809529/go.mod h1:O
 github.com/pingcap/tidb-dashboard v0.0.0-20211107164327-80363dfbe884/go.mod h1:OCXbZTBTIMRcIt0jFsuCakZP+goYRv6IjawKbwLS2TQ=
 github.com/pingcap/tidb-tools v5.2.2-0.20211019062242-37a8bef2fa17+incompatible h1:c7+izmker91NkjkZ6FgTlmD4k1A5FLOAq+li6Ki2/GY=
 github.com/pingcap/tidb-tools v5.2.2-0.20211019062242-37a8bef2fa17+incompatible/go.mod h1:XGdcy9+yqlDSEMTpOXnwf3hiTeqrV6MN/u1se9N8yIM=
-github.com/pingcap/tipb v0.0.0-20211116093845-e9b045a0bdf8 h1:Vu/6oq8EFNWgyXRHiclNzTKIu+YKHPCSI/Ba5oVrLtM=
-github.com/pingcap/tipb v0.0.0-20211116093845-e9b045a0bdf8/go.mod h1:A7mrd7WHBl1o63LE2bIBGEJMTNWXqhgmYiOvMLxozfs=
+github.com/pingcap/tipb v0.0.0-20211201080053-bd104bb270ba h1:Tt5W/maVBUbG+wxg2nfc88Cqj/HiWYb0TJQ2Rfi0UOQ=
+github.com/pingcap/tipb v0.0.0-20211201080053-bd104bb270ba/go.mod h1:A7mrd7WHBl1o63LE2bIBGEJMTNWXqhgmYiOvMLxozfs=
 github.com/pkg/errors v0.8.0/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pkg/errors v0.8.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=

parser/ast/functions.go

@@ -279,6 +279,7 @@ const (
 	IsIPv4Mapped    = "is_ipv4_mapped"
 	IsIPv6          = "is_ipv6"
 	IsUsedLock      = "is_used_lock"
+	IsUUID          = "is_uuid"
 	MasterPosWait   = "master_pos_wait"
 	NameConst       = "name_const"
 	ReleaseAllLocks = "release_all_locks"