使用flex和bison实现的sql引擎解析

因为老师要求，近期在做oceanbase存储过程的实现，在oceanbase 0.4曾经是不支持存储过程的。实现的主要步骤主要包含

1、语法解析

2、词法解析

3、详细运行语法树的步骤

如今先来说说语法解析吧，在这一块主要是使用的flex（ 词法分析器生成工具） 和bison(语法分析器生成器) 这两个是对用户输入的存储过程语句进行解析的

来详细说说该怎么实现对sql语句的分析吧

1、首先建立一个lex的文件

%option noyywrap nodefault yylineno case-insensitive

%{

#include "prosql.tab.hpp"
#include <stdarg.h>
#include <string.h>
#include <stdlib.h>
#include <stdio.h>
#include <malloc.h>
//YYSTYPE yylval;
int oldstate;
extern "C" int yylex();
//extern "C" int yyparse();
extern "C" void yyerror(const char *s, ...);
extern char globalInputText[10000];
extern int readInputForLexer( char *buffer, int *numBytesRead, int maxBytesToRead );
#undef YY_INPUT
#define YY_INPUT(b,r,s) readInputForLexer(b,&r,s)
%}

%x COMMENT

%%

CREATE		{ return CREATE; }
PROCEDURE	{ return PROCEDURE; }
SQL		{ return SQL; }

DECLARE		{ return DECLARE; }
SET		{ return SET; }
BEGIN		{ return BEGINT; }
END		{ return END; }

INT		{ return INT; }
VARCHAR		{ return VARCHAR; }
DATE		{ return DATE; }
TIME		{ return TIME; }
DOUBLE		{ return DOUBLE; }

IF		{ return IF; }
THEN		{ return THEN; }
ELSE		{ return ELSE; }
ENDIF		{ return ENDIF; }
FOR		{ return FOR; }
WHEN		{ return WHEN; }
WHILE		{ return WHILE; }


[0-9]+	{ yylval.strval = strdup(yytext);/*printf("number=%s
",yylval.strval);*/ return INTNUM; }/*number*/

[0-9]+"."[0-9]* |
"."[0-9]+	|
[0-9]+E[-+]?[0-9]+	|
[0-9]+"."[0-9]*E[-+]?[0-9]+ |
"."[0-9]*E[-+]?[0-9]+	{ yylval.strval = strdup(yytext);/*printf("float=%s
",yylval.strval);*/ return APPROXNUM; }/*double*/

TRUE	{ yylval.strval = "1";/*printf("bool=%s
",yylval.strval);*/ return BOOL; }/*bool*/

FALSE	{ yylval.strval = "0";/*printf("bool=%s
",yylval.strval);*/ return BOOL; }/*bool*/

'(\.|''|[^'
])*'	|
"(\.|""|[^"
])*"  {
				char *temp = strdup(yytext); 
				yylval.strval = strdup(yytext);

				//GetCorrectString(yylval.strval, temp);
				
				/*printf("string=%s
",yylval.strval);*/
				return STRING;
			}/*string*/
'(\.|[^'
])*$		{ yyerror("Unterminated string %s", yytext); }
"(\.|[^"
])*$		{ yyerror("Unterminated string %s", yytext); }


X'[0-9A-F]+' |  
0X[0-9A-F]+  	{ yylval.strval = strdup(yytext); return STRING; }


0B[01]+      |
B'[01]+'     { yylval.strval = strdup(yytext); return STRING; }


[-+&~|^/%*(),.;!]   { return yytext[0]; }

"&&"	{ return ANDOP; }
"||"	{ return OR; }

"<"	{ yylval.subtok = 1; return COMPARISON; }
">"	{ yylval.subtok = 2; return COMPARISON; }
"!="	|
"<>"	{ yylval.subtok = 3; return COMPARISON; }
"="	{ yylval.subtok = 4; return COMPARISON; }
"<="	{ yylval.subtok = 5; return COMPARISON; }
">="	{ yylval.subtok = 6; return COMPARISON; }
"<=>"	{ yylval.subtok = 12; return COMPARISON; }

"<<"	{ yylval.subtok = 1; return SHIFT; }
">>"	{ yylval.subtok = 2; return SHIFT; }


[A-Za-z][A-Za-z0-9_]*	{ yylval.strval = strdup(yytext);
			  /*printf("name 1=%s
",yylval.strval);*/
                          return NAME; }
`[^`/\.
]+`           { yylval.strval = strdup(yytext+1);
			  /*printf("name 2=%s
",yylval.strval);*/
                          yylval.strval[yyleng-2] = 0;
                          return NAME; }

`[^`
]*$               { yyerror("unterminated quoted name %s", yytext); }


@[0-9a-z_.$]+ |
@"[^"
]+" |
@`[^`
]+` |
@'[^'
]+' { yylval.strval = strdup(yytext+1);  return USERVAR; }

@"[^"
]*$ { yyerror("unterminated quoted user variable %s", yytext); }
@`[^`
]*$ { yyerror("unterminated quoted user variable %s", yytext); }
@'[^'
]*$ { yyerror("unterminated quoted user variable %s", yytext); }

":="     { return ASSIGN; }

#.*		;
"--"[ 	].*	;

"/*"            { oldstate = YY_START; BEGIN COMMENT; }
<COMMENT>"*/"   { BEGIN oldstate; }
<COMMENT>.|
   ;
<COMMENT><<EOF>> { yyerror("unclosed comment"); }


[ 	
]         /* white space */
.               { yyerror("mystery character '%c'", *yytext); }

%%

这一部分呢就是对 每一个我们自己定义的满足正则的识别

接下来是对词的语法识别

%{
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
#include <stdio.h>
#include <malloc.h>
char * parsetreeroot=NULL;
extern "C" int yylex();
extern "C" int yyparse();
extern "C" void yyerror(const char *s, ...);
char globalInputText[10000];
int globalReadOffset;
int readInputForLexer( char *buffer, int *numBytesRead, int maxBytesToRead );
char * mystrcat(char *s1,char *s2)
{
	char *p1=(char *)malloc(strlen(s1)+strlen(s2)+1);
	strcpy(p1,s1);
	strcat(p1,s2);
	return p1;
}
%}
%locations
%union {
	int intval;		
	double floatval;	
	char *strval;
	int subtok;
}
%token <strval> NAME
%token <strval> STRING
%token <strval> INTNUM
%token <strval> BOOL
%token <strval> APPROXNUM
%token <strval> USERVAR

%type <strval> stmt_root  create_stmt para_list definition  data_type pro_block pro_parameters declare_list set_list 
%type <strval> assign_var  pro_body pro_stmt_list sql_stmt expr



%right ASSIGN
%left OR
%left XOR
%left ANDOP

%left NOT '!'
%left BETWEEN
%left <subtok> COMPARISON /* = <> < > <= >= <=> */
%left '|'
%left '&'
%left <subtok> SHIFT /* << >> */
%left '+' '-'
%left '*' '/' '%' MOD
%left '^'

%token CREATE
%token PROCEDURE
%token PRONAME
%token DECLARE
%token SET
%token BEGINT
%token END
%token SQL

%token INT
%token VARCHAR
%token DATE
%token TIME
%token DOUBLE

%token IF
%token NOT
%token EXISTS
%token THEN
%token ELSE
%token ENDIF
%token FOR
%token WHEN
%token WHILE
%start stmt_root
%%

stmt_root: create_stmt pro_block { $$=mystrcat($1,$2); parsetreeroot=$$;}
;
create_stmt: CREATE PROCEDURE  NAME '(' para_list ')' 
		{
			char *temp=mystrcat("create procedure ",$3);
			temp=mystrcat(temp,"(");
			temp=mystrcat(temp,$5);
			$$=mystrcat(temp,")(create)
");
		}
;
/*
opt_if_not_exists:	      { $$ = 0; } 
   | IF NOT EXISTS            { $$ = 1; } 
   ;
*/
para_list: definition { $$=$1; }
|definition ',' para_list 
		{	
			
			char *temp=mystrcat($1,",");
			$$=mystrcat(temp,$3); 
		}
;
definition: USERVAR data_type 
		{	
			
			char *temp=mystrcat($1," ");
			$$=mystrcat(temp,$2);
			
		}
;

data_type:
   DATE 					{$$="date"; }
   | TIME					{$$="time"; }
   | VARCHAR '(' INTNUM ')' 			{$$="varchar"; }
   | INT 					{$$="int"; }
   | DOUBLE 					{$$="double"; }	
   ;

pro_block: BEGINT pro_parameters pro_body END 
		{ 
			char *temp=mystrcat("begin
",$2);
			temp=mystrcat(temp,"");
			temp=mystrcat(temp,$3);
			$$=mystrcat(temp,"end");
			//printf("pro_body %s
",$3);
		}
;

pro_parameters: declare_list ';' { $$=mystrcat($1,";(declare)
");}
|pro_parameters  declare_list ';' 
		{ 
			char *temp=mystrcat($1,$2);
			$$=mystrcat(temp,";(declare)
");
		}
|pro_parameters  set_list ';' 
		{
	 		char *temp=mystrcat($1,$2);
			$$=mystrcat(temp,";(set)
");
		}
;

declare_list:
|DECLARE definition 
		{
			$$=mystrcat("declare ",$2);
		}
|declare_list ',' definition 
		{ 
			char *temp=mystrcat($1,",");
			$$=mystrcat(temp,$3); 
		}
;

set_list:
|SET assign_var 
		{
			$$=mystrcat("set ",$2); 
		}
| set_list ',' assign_var 
		{	
			char *temp=mystrcat($1,",");
			$$=mystrcat(temp,$3); 
		}
;

assign_var : USERVAR COMPARISON expr
		{ 	
			char *temp=mystrcat($1,"=");
			$$=mystrcat(temp,$3); 
		}
;

expr: NAME         { $$=$1;}
   | STRING        { $$=$1;}
   | INTNUM        { $$=$1;}
   | APPROXNUM 	   { $$=$1;}	
   | BOOL          { $$=$1;}
   ;

pro_body :  pro_stmt_list { $$=$1; }
;
pro_stmt_list: sql_stmt {$$=$1; }
|pro_stmt_list  sql_stmt 
		{
			$$=mystrcat($1,$2);
		}
;
sql_stmt: 
|SQL NAME ';' { $$=mystrcat($2,";(sql)
");}
;
%%
/*
int main(int argc, char* argv[])
{
	yyparse();
}*/
int readInputForLexer( char *buffer, int *numBytesRead, int maxBytesToRead ) {
	int numBytesToRead = maxBytesToRead;
	int bytesRemaining = strlen(globalInputText)-globalReadOffset;
	int i;
	if ( numBytesToRead > bytesRemaining ) { numBytesToRead = bytesRemaining; }
	for ( i = 0; i < numBytesToRead; i++ ) {
		buffer[i] = globalInputText[globalReadOffset+i];
	}
	*numBytesRead = numBytesToRead;
	globalReadOffset += numBytesToRead;
	return 0;
}
void yyerror(const char *s, ...)
{
    fprintf(stderr, "error: %s
", s);
}
void zzerror(const char *s, ...)
{
	extern int  yylineno;

	va_list ap;
	va_start(ap, s);

	fprintf(stderr, "%d: error: ", yylineno);
	vfprintf(stderr, s, ap);
	fprintf(stderr, "
");
}

int yywrap(void)
{
    return 1;
}
char* getsql()
{	
	return parsetreeroot;
}

 

这部分就是对上一个识别出来的词 进行顺序上的确定，构成一个完整的语法

这些须要在linux环境下进行调试

bison -d 文件名称

flex 文件名称