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.

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

expression: set correct return field type of builtin CONVERT() (#9305) #10263

Merged
merged 4 commits into from
May 23, 2019
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
10 changes: 8 additions & 2 deletions ddl/db_integration_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -114,7 +114,7 @@ func (s *testIntegrationSuite) TestInvalidDefault(c *C) {
c.Assert(terror.ErrorEqual(err, types.ErrInvalidDefault), IsTrue, Commentf("err %v", err))
}

// for issue #3848
// TestInvalidNameWhenCreateTable for issue #3848
func (s *testIntegrationSuite) TestInvalidNameWhenCreateTable(c *C) {
tk := testkit.NewTestKit(c, s.store)

Expand All @@ -133,7 +133,7 @@ func (s *testIntegrationSuite) TestInvalidNameWhenCreateTable(c *C) {
c.Assert(terror.ErrorEqual(err, ddl.ErrWrongDBName), IsTrue, Commentf("err %v", err))
}

// for issue #6879
// TestCreateTableIfNotExists for issue #6879
func (s *testIntegrationSuite) TestCreateTableIfNotExists(c *C) {
tk := testkit.NewTestKit(c, s.store)

Expand Down Expand Up @@ -269,6 +269,12 @@ func (s *testIntegrationSuite) TestChangingTableCharset(c *C) {
rs.Close()
}
c.Assert(err.Error(), Equals, "[parser:1115]Unknown character set: 'gbk'")
rs, err = tk.Exec("alter table t charset utf8")
if rs != nil {
rs.Close()
}
c.Assert(err.Error(), Equals, "[ddl:210]unsupported modify charset from latin1 to utf8")

rs, err = tk.Exec("alter table t charset utf8 collate latin1_bin")
if rs != nil {
rs.Close()
Expand Down
44 changes: 34 additions & 10 deletions expression/builtin_string.go
Original file line number Diff line number Diff line change
Expand Up @@ -924,9 +924,33 @@ func (c *convertFunctionClass) getFunction(ctx sessionctx.Context, args []Expres
return nil, errors.Trace(err)
}
bf := newBaseBuiltinFuncWithTp(ctx, args, types.ETString, types.ETString, types.ETString)
// TODO: issue #4436: The second parameter should be a constant.
// TODO: issue #4474: Charset supported by TiDB and MySQL is not the same.
// TODO: Fix #4436 && #4474, set the correct charset and flag of `bf.tp`.

charsetArg, ok := args[1].(*Constant)
if !ok {
// `args[1]` is limited by parser to be a constant string,
// should never go into here.
return nil, errIncorrectArgs.GenWithStackByArgs("charset")
}
transcodingName := charsetArg.Value.GetString()
bf.tp.Charset = strings.ToLower(transcodingName)
// Quoted about the behavior of syntax CONVERT(expr, type) to CHAR():
// In all cases, the string has the default collation for the character set.
// See https://dev.mysql.com/doc/refman/5.7/en/cast-functions.html#function_convert
// Here in syntax CONVERT(expr USING transcoding_name), behavior is kept the same,
// picking the default collation of target charset.
var err error
bf.tp.Collate, err = charset.GetDefaultCollation(bf.tp.Charset)
if err != nil {
return nil, errUnknownCharacterSet.GenWithStackByArgs(transcodingName)
}
// Result will be a binary string if converts charset to BINARY.
// See https://dev.mysql.com/doc/refman/5.7/en/charset-binary-set.html
if types.IsBinaryStr(bf.tp) {
types.SetBinChsClnFlag(bf.tp)
} else {
bf.tp.Flag &= ^mysql.BinaryFlag
}

bf.tp.Flen = mysql.MaxBlobWidth
sig := &builtinConvertSig{bf}
return sig, nil
Expand All @@ -943,21 +967,21 @@ func (b *builtinConvertSig) Clone() builtinFunc {
}

// evalString evals CONVERT(expr USING transcoding_name).
// Syntax CONVERT(expr, type) is parsed as cast expr so not handled here.
// See https://dev.mysql.com/doc/refman/5.7/en/cast-functions.html#function_convert
func (b *builtinConvertSig) evalString(row chunk.Row) (string, bool, error) {
expr, isNull, err := b.args[0].EvalString(b.ctx, row)
if isNull || err != nil {
return "", true, errors.Trace(err)
}

charsetName, isNull, err := b.args[1].EvalString(b.ctx, row)
if isNull || err != nil {
return "", true, errors.Trace(err)
}

encoding, _ := charset.Lookup(charsetName)
// 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 "", true, errUnknownCharacterSet.GenWithStackByArgs(charsetName)
return "", true, errUnknownCharacterSet.GenWithStackByArgs(b.tp.Charset)
}

target, _, err := transform.String(encoding.NewDecoder(), expr)
Expand Down
57 changes: 42 additions & 15 deletions expression/builtin_string_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -676,41 +676,61 @@ func (s *testEvaluatorSuite) TestSubstring(c *C) {
func (s *testEvaluatorSuite) TestConvert(c *C) {
defer testleak.AfterTest(c)()
tbl := []struct {
str string
cs string
result string
str interface{}
cs string
result string
hasBinaryFlag bool
}{
{"haha", "utf8", "haha"},
{"haha", "ascii", "haha"},
{"haha", "binary", "haha"},
{"haha", "utf8", "haha", false},
{"haha", "ascii", "haha", false},
{"haha", "binary", "haha", true},
{"haha", "bInAry", "haha", true},
{types.NewBinaryLiteralFromUint(0x7e, -1), "BiNarY", "~", true},
{types.NewBinaryLiteralFromUint(0xe4b8ade696870a, -1), "uTf8", "中文\n", false},
}
for _, v := range tbl {
fc := funcs[ast.Convert]
f, err := fc.getFunction(s.ctx, s.datumsToConstants(types.MakeDatums(v.str, v.cs)))
c.Assert(err, IsNil)
c.Assert(f, NotNil)
retType := f.getRetTp()
c.Assert(retType.Charset, Equals, strings.ToLower(v.cs))
collate, err := charset.GetDefaultCollation(strings.ToLower(v.cs))
c.Assert(err, IsNil)
c.Assert(retType.Collate, Equals, collate)
c.Assert(mysql.HasBinaryFlag(retType.Flag), Equals, v.hasBinaryFlag)

r, err := evalBuiltinFunc(f, chunk.Row{})
c.Assert(err, IsNil)
c.Assert(r.Kind(), Equals, types.KindString)
c.Assert(r.GetString(), Equals, v.result)
}

// Test case for error
// Test case for getFunction() error
errTbl := []struct {
str interface{}
cs string
result string
str interface{}
cs string
err string
}{
{"haha", "wrongcharset", "haha"},
{"haha", "wrongcharset", "[expression:1115]Unknown character set: 'wrongcharset'"},
{"haha", "cp866", "[expression:1115]Unknown character set: 'cp866'"},
}
for _, v := range errTbl {
fc := funcs[ast.Convert]
f, err := fc.getFunction(s.ctx, s.datumsToConstants(types.MakeDatums(v.str, v.cs)))
c.Assert(err, IsNil)
c.Assert(f, NotNil)
_, err = evalBuiltinFunc(f, chunk.Row{})
c.Assert(err, NotNil)
c.Assert(err.Error(), Equals, v.err)
c.Assert(f, IsNil)
}

// Test wrong charset while evaluating.
fc := funcs[ast.Convert]
f, err := fc.getFunction(s.ctx, s.datumsToConstants(types.MakeDatums("haha", "utf8")))
c.Assert(err, IsNil)
c.Assert(f, NotNil)
wrongFunction := f.(*builtinConvertSig)
wrongFunction.tp.Charset = "wrongcharset"
_, err = evalBuiltinFunc(wrongFunction, chunk.Row{})
c.Assert(err.Error(), Equals, "[expression:1115]Unknown character set: 'wrongcharset'")
}

func (s *testEvaluatorSuite) TestSubstringIndex(c *C) {
Expand Down Expand Up @@ -1199,6 +1219,13 @@ func (s *testEvaluatorSuite) TestChar(c *C) {
r, err := evalBuiltinFunc(f, chunk.Row{})
c.Assert(err, IsNil)
c.Assert(r, testutil.DatumEquals, types.NewDatum("AB"))

// Test unsupported charset.
fc = funcs[ast.CharFunc]
f, err = fc.getFunction(s.ctx, s.datumsToConstants(types.MakeDatums("65", "tidb")))
c.Assert(err, IsNil)
_, err = evalBuiltinFunc(f, chunk.Row{})
c.Assert(err.Error(), Equals, "unknown encoding: tidb")
}

func (s *testEvaluatorSuite) TestCharLength(c *C) {
Expand Down
13 changes: 11 additions & 2 deletions expression/integration_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -558,6 +558,7 @@ func (s *testIntegrationSuite) TestStringBuiltin(c *C) {
tk := testkit.NewTestKit(c, s.store)
tk.MustExec("use test")
ctx := context.Background()
var err error

// for length
tk.MustExec("drop table if exists t")
Expand Down Expand Up @@ -899,8 +900,16 @@ func (s *testIntegrationSuite) TestStringBuiltin(c *C) {
result.Check(testkit.Rows("'121' '0' '中文' <nil>"))

// for convert
result = tk.MustQuery(`select convert("123" using "binary"), convert("123" using "binary"), convert("中文" using "binary"), convert("中文" using "utf8"), convert("中文" using "utf8mb4"), convert(cast("中文" as binary) using "utf8");`)
result.Check(testkit.Rows("123 123 中文 中文 中文 中文"))
result = tk.MustQuery(`select convert("123" using "binary"), convert("中文" using "binary"), convert("中文" using "utf8"), convert("中文" using "utf8mb4"), convert(cast("中文" as binary) using "utf8");`)
result.Check(testkit.Rows("123 中文 中文 中文 中文"))
// Charset 866 does not have a default collation configured currently, so this will return error.
_, err = tk.Exec(`select convert("123" using "866");`)
c.Assert(err.Error(), Equals, "[parser:1115]Unknown character set: '866'")
// Test case in issue #4436.
tk.MustExec("drop table if exists t;")
tk.MustExec("create table t(a char(20));")
_, err = tk.Exec("select convert(a using a) from t;")
c.Assert(err.Error(), Equals, "[parser:1115]Unknown character set: 'a'")

// for insert
result = tk.MustQuery(`select insert("中文", 1, 1, cast("aaa" as binary)), insert("ba", -1, 1, "aaa"), insert("ba", 1, 100, "aaa"), insert("ba", 100, 1, "aaa");`)
Expand Down
2 changes: 2 additions & 0 deletions expression/typeinfer_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -459,6 +459,8 @@ func (s *testInferTypeSuite) createTestCase4StrFuncs() []typeInferTestCase {
{"convert(c_binary using utf8mb4)", mysql.TypeLongBlob, charset.CharsetUTF8MB4, 0, mysql.MaxBlobWidth, types.UnspecifiedLength},
{"convert(c_binary using utf8mb4)", mysql.TypeLongBlob, charset.CharsetUTF8MB4, 0, mysql.MaxBlobWidth, types.UnspecifiedLength},
{"convert(c_text_d using utf8mb4)", mysql.TypeLongBlob, charset.CharsetUTF8MB4, 0, mysql.MaxBlobWidth, types.UnspecifiedLength},
{"convert(c_binary using 'binary')", mysql.TypeLongBlob, charset.CharsetBin, mysql.BinaryFlag, mysql.MaxBlobWidth, types.UnspecifiedLength},
{"convert(c_text_d using 'binary')", mysql.TypeLongBlob, charset.CharsetBin, mysql.BinaryFlag, mysql.MaxBlobWidth, types.UnspecifiedLength},

{"insert(c_varchar, c_int_d, c_int_d, c_varchar)", mysql.TypeLongBlob, charset.CharsetUTF8MB4, 0, mysql.MaxBlobWidth, types.UnspecifiedLength},
{"insert(c_varchar, c_int_d, c_int_d, c_binary)", mysql.TypeLongBlob, charset.CharsetBin, mysql.BinaryFlag, mysql.MaxBlobWidth, types.UnspecifiedLength},
Expand Down
Loading