feat: MySQL数据源支持基于识别自定义分隔符以及识别BeginEnd语句块切分SQL文本

go.mod

@@ -185,7 +185,7 @@ replace (
 	cloud.google.com/go/compute/metadata => cloud.google.com/go/compute/metadata v0.1.0
 	github.com/labstack/echo/v4 => github.com/labstack/echo/v4 v4.6.1
 	github.com/pingcap/log => github.com/pingcap/log v0.0.0-20191012051959-b742a5d432e9
-	github.com/pingcap/parser => github.com/sjjian/parser v0.0.0-20240305095250-688ad439ef31
+	github.com/pingcap/parser => github.com/sjjian/parser v0.0.0-20240704052347-b6199b7bccae
 	github.com/swaggo/swag => github.com/swaggo/swag v1.6.7
 	google.golang.org/grpc => google.golang.org/grpc v1.29.0
 )

go.sum

@@ -795,8 +795,8 @@ github.com/sirupsen/logrus v1.6.0/go.mod h1:7uNnSEd1DgxDLC74fIahvMZmmYsHGZGEOFrf
 github.com/sirupsen/logrus v1.7.0/go.mod h1:yWOB1SBYBC5VeMP7gHvWumXLIWorT60ONWic61uBYv0=
 github.com/sirupsen/logrus v1.9.0 h1:trlNQbNUG3OdDrDil03MCb1H2o9nJ1x4/5LYw7byDE0=
 github.com/sirupsen/logrus v1.9.0/go.mod h1:naHLuLoDiP4jHNo9R0sCBMtWGeIprob74mVsIT4qYEQ=
-github.com/sjjian/parser v0.0.0-20240305095250-688ad439ef31 h1:E3JSX1FjUlg8ep8XQlLkdTFbZhM4tmecpdfZhUuubLs=
-github.com/sjjian/parser v0.0.0-20240305095250-688ad439ef31/go.mod h1:Qq2tnreUXwVo7NAKAHmbWFsgqpDUkxwhJCClY+ZCudA=
+github.com/sjjian/parser v0.0.0-20240704052347-b6199b7bccae h1:rTwogc7Uq0/7zMHNhpQTqjtGKdccVh8Z8vHReIlsXG0=
+github.com/sjjian/parser v0.0.0-20240704052347-b6199b7bccae/go.mod h1:Qq2tnreUXwVo7NAKAHmbWFsgqpDUkxwhJCClY+ZCudA=
 github.com/skeema/knownhosts v1.2.0 h1:h9r9cf0+u7wSE+M183ZtMGgOJKiL96brpaz5ekfJCpM=
 github.com/skeema/knownhosts v1.2.0/go.mod h1:g4fPeYpque7P0xefxtGzV81ihjC8sX2IqpAoNkjxbMo=
 github.com/smartystreets/assertions v0.0.0-20180927180507-b2de0cb4f26d/go.mod h1:OnSkiWE9lh6wB0YB77sQom3nweQdgAjqCqsofrRNTgc=

sqle/driver/mysql/splitter/block.go

@@ -0,0 +1,80 @@
+package splitter
+
+import (
+	"github.com/pingcap/parser"
+	"strings"
+)
+
+type Block interface {
+	MatchBegin(token *parser.Token) bool
+	MatchEnd(token *parser.Token) bool
+}
+
+var allBlocks []Block = []Block{
+	BeginEndBlock{},
+	IfEndIfBlock{},
+	CaseEndCaseBlock{},
+	RepeatEndRepeatBlock{},
+	WhileEndWhileBlock{},
+	LoopEndLoopBlock{},
+}
+
+type LoopEndLoopBlock struct{}
+
+func (b BeginEndBlock) MatchBegin(token *parser.Token) bool {
+	return token.TokenType() == parser.Begin
+}
+
+func (b BeginEndBlock) MatchEnd(token *parser.Token) bool {
+	return true
+}
+
+type IfEndIfBlock struct{}
+
+func (b IfEndIfBlock) MatchBegin(token *parser.Token) bool {
+	return token.TokenType() == parser.IfKwd
+}
+
+func (b IfEndIfBlock) MatchEnd(token *parser.Token) bool {
+	return token.TokenType() == parser.IfKwd
+}
+
+type CaseEndCaseBlock struct{}
+
+func (b CaseEndCaseBlock) MatchBegin(token *parser.Token) bool {
+	return token.TokenType() == parser.CaseKwd
+}
+
+func (b CaseEndCaseBlock) MatchEnd(token *parser.Token) bool {
+	return token.TokenType() == parser.CaseKwd
+}
+
+type RepeatEndRepeatBlock struct{}
+
+func (b RepeatEndRepeatBlock) MatchBegin(token *parser.Token) bool {
+	return token.TokenType() == parser.Repeat
+}
+
+func (b RepeatEndRepeatBlock) MatchEnd(token *parser.Token) bool {
+	return token.TokenType() == parser.Repeat
+}
+
+type WhileEndWhileBlock struct{}
+
+func (b WhileEndWhileBlock) MatchBegin(token *parser.Token) bool {
+	return token.TokenType() == parser.Identifier && strings.ToUpper(token.Ident()) == "WHILE"
+}
+
+func (b WhileEndWhileBlock) MatchEnd(token *parser.Token) bool {
+	return token.TokenType() == parser.Identifier && strings.ToUpper(token.Ident()) == "WHILE"
+}
+
+type BeginEndBlock struct{}
+
+func (b LoopEndLoopBlock) MatchBegin(token *parser.Token) bool {
+	return token.TokenType() == parser.Identifier && strings.ToUpper(token.Ident()) == "LOOP"
+}
+
+func (b LoopEndLoopBlock) MatchEnd(token *parser.Token) bool {
+	return token.TokenType() == parser.Identifier && strings.ToUpper(token.Ident()) == "LOOP"
+}

sqle/driver/mysql/splitter/delimiter.go

@@ -0,0 +1,138 @@
+package splitter
+
+import (
+	"errors"
+	"github.com/pingcap/parser"
+	"strings"
+)
+
+const (
+	BackSlash              int    = '\\'
+	BackSlashString        string = "\\"
+	BlankSpace             string = " "
+	DefaultDelimiterString string = ";"
+	DelimiterCommand       string = "DELIMITER"
+	DelimiterCommandSort   string = `\d`
+)
+
+type Delimiter struct {
+	FirstTokenTypeOfDelimiter  int
+	FirstTokenValueOfDelimiter string
+	DelimiterStr               string
+	line                       int
+	startPos                   int
+}
+
+func NewDelimiter() *Delimiter {
+	return &Delimiter{}
+}
+
+// \\d会被识别为三个token \ \ d 不能使用Lex，Lex可能会跳过空格和注释，因此这里使用字符串匹配
+func (d *Delimiter) isSortDelimiterCommand(sql string, index int) bool {
+	return index+2 < len(sql) && sql[index+1] == 'd'
+}
+
+// DELIMITER会被识别为identifier，因此这里仅需识别其值是否相等
+func (d *Delimiter) isDelimiterCommand(token string) bool {
+	return strings.ToUpper(token) == DelimiterCommand
+}
+
+// 该函数翻译自MySQL Client获取delimiter值的代码，参考：https://github.com/mysql/mysql-server/blob/824e2b4064053f7daf17d7f3f84b7a3ed92e5fb4/client/mysql.cc#L4866
+func getDelimiter(line string) string {
+	ptr := 0
+	start := 0
+	quoted := false
+	qtype := byte(0)
+
+	// 跳过开头的空格
+	for ptr < len(line) && isSpace(line[ptr]) {
+		ptr++
+	}
+
+	if ptr == len(line) {
+		return ""
+	}
+
+	// 检查是否为引号字符串
+	if line[ptr] == '\'' || line[ptr] == '"' || line[ptr] == '`' {
+		qtype = line[ptr]
+		quoted = true
+		ptr++
+	}
+
+	start = ptr
+
+	// 找到字符串结尾
+	for ptr < len(line) {
+		if !quoted && line[ptr] == '\\' && ptr+1 < len(line) { // 跳过转义字符
+			ptr += 2
+		} else if (!quoted && isSpace(line[ptr])) || (quoted && line[ptr] == qtype) {
+			break
+		} else {
+			ptr++
+		}
+	}
+
+	return line[start:ptr]
+}
+
+// 辅助函数,判断字符是否为空格
+func isSpace(c byte) bool {
+	return c == ' ' || c == '\t' || c == '\n' || c == '\r'
+}
+
+var ErrDelimiterCanNotExtractToken = errors.New("sorry, we cannot extract any token form the delimiter you provide, please change a delimiter")
+var ErrDelimiterContainsBackslash = errors.New("DELIMITER cannot contain a backslash character")
+var ErrDelimiterContainsBlankSpace = errors.New("DELIMITER should not contain blank space")
+var ErrDelimiterMissing = errors.New("DELIMITER must be followed by a 'delimiter' character or string")
+var ErrDelimiterReservedKeyword = errors.New("delimiter should not use a reserved keyword")
+
+/*
+该方法设置分隔符，对分隔符的内容有一定的限制：
+
+ 1. 不允许分隔符内部包含反斜杠
+ 2. 不允许分隔符为空字符串
+ 3. 不允许分隔符为mysql的保留字，因为这样会被scanner扫描为其他类型的token，从而绕过判断分隔符的逻辑
+
+注：其中1和2与MySQL客户端对分隔符内容一致，错误内容参考MySQL客户端源码中的com_delimiter函数
+https://github.com/mysql/mysql-server/blob/824e2b4064053f7daf17d7f3f84b7a3ed92e5fb4/client/mysql.cc#L4621
+*/
+func (d *Delimiter) setDelimiter(delimiter string) (err error) {
+	if delimiter == "" {
+		return ErrDelimiterMissing
+	}
+	if strings.Contains(delimiter, BackSlashString) {
+		return ErrDelimiterContainsBackslash
+	}
+	if strings.Contains(delimiter, BlankSpace) {
+		return ErrDelimiterContainsBlankSpace
+	}
+	if isReservedKeyWord(delimiter) {
+		return ErrDelimiterReservedKeyword
+	}
+	token := parser.NewScanner(delimiter).NextToken()
+	d.FirstTokenTypeOfDelimiter = token.TokenType()
+	if d.FirstTokenTypeOfDelimiter == 0 {
+		return ErrDelimiterCanNotExtractToken
+	}
+	d.FirstTokenValueOfDelimiter = token.Ident()
+	d.DelimiterStr = delimiter
+	return nil
+}
+
+func isReservedKeyWord(input string) bool {
+	token := parser.NewScanner(input).NextToken()
+	tokenType := token.TokenType()
+	if len(token.Ident()) < len(input) {
+		// 如果分隔符无法识别为一个token，则一定不是关键字
+		return false
+	}
+	// 如果分隔符识别为一个关键字，但不知道是哪个关键字，则为identifier，此时就非保留字
+	return tokenType != parser.Identifier && tokenType > parser.YyEOFCode && tokenType < parser.YyDefault
+}
+
+func (d *Delimiter) reset() error {
+	d.line = 0
+	d.startPos = 0
+	return d.setDelimiter(DefaultDelimiterString)
+}