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\n",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\n",yylval.strval);*/ return APPROXNUM; }/*double*/TRUE  { yylval.strval = "1";/*printf("bool=%s\n",yylval.strval);*/ return BOOL; }/*bool*/FALSE  { yylval.strval = "0";/*printf("bool=%s\n",yylval.strval);*/ return BOOL; }/*bool*/'(\\.|''|[^'\n])*'    |
\"(\\.|\"\"|[^"\n])*\"  {char *temp = strdup(yytext); yylval.strval = strdup(yytext);//GetCorrectString(yylval.strval, temp);/*printf("string=%s\n",yylval.strval);*/return STRING;}/*string*/
'(\\.|[^'\n])*$       { yyerror("Unterminated string %s", yytext); }
\"(\\.|[^"\n])*$      { 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\n",yylval.strval);*/return NAME; }
`[^`/\\.\n]+`           { yylval.strval = strdup(yytext+1);/*printf("name 2=%s\n",yylval.strval);*/yylval.strval[yyleng-2] = 0;return NAME; }`[^`\n]*$               { yyerror("unterminated quoted name %s", yytext); }@[0-9a-z_.$]+ |
@\"[^"\n]+\" |
@`[^`\n]+` |
@'[^'\n]+' { yylval.strval = strdup(yytext+1);  return USERVAR; }@\"[^"\n]*$ { yyerror("unterminated quoted user variable %s", yytext); }
@`[^`\n]*$ { yyerror("unterminated quoted user variable %s", yytext); }
@'[^'\n]*$ { yyerror("unterminated quoted user variable %s", yytext); }":="     { return ASSIGN; }#.*       ;
"--"[ \t].*   ;"/*"            { oldstate = YY_START; BEGIN COMMENT; }
<COMMENT>"*/"   { BEGIN oldstate; }
<COMMENT>.|\n   ;
<COMMENT><<EOF>> { yyerror("unclosed comment"); }[ \t\n]         /* 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);$$=mystrcat(temp,")(create)\n");}
;
/*
opt_if_not_exists:        { $$ = 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                     {$$="double"; }  ;pro_block: BEGINT pro_parameters pro_body END { char *temp=mystrcat("begin\n",$2);temp=mystrcat(temp,"");temp=mystrcat(temp,$3);$$=mystrcat(temp,"end");//printf("pro_body %s\n",$3);}
;pro_parameters: declare_list ';' { $$=mystrcat($1,";(declare)\n");}
|pro_parameters  declare_list ';' { char *temp=mystrcat($1,$2);$$=mystrcat(temp,";(declare)\n");}
|pro_parameters  set_list ';' {char *temp=mystrcat($1,$2);$$=mystrcat(temp,";(set)\n");}
;declare_list:
|DECLARE definition {$$=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)\n");}
;
%%
/*
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\n", 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, "\n");
}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