xref: /sqlite-3.40.0/src/tokenize.c (revision 7c68d60b)
1 /*
2 ** Copyright (c) 1999, 2000 D. Richard Hipp
3 **
4 ** This program is free software; you can redistribute it and/or
5 ** modify it under the terms of the GNU General Public
6 ** License as published by the Free Software Foundation; either
7 ** version 2 of the License, or (at your option) any later version.
8 **
9 ** This program is distributed in the hope that it will be useful,
10 ** but WITHOUT ANY WARRANTY; without even the implied warranty of
11 ** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
12 ** General Public License for more details.
13 **
14 ** You should have received a copy of the GNU General Public
15 ** License along with this library; if not, write to the
16 ** Free Software Foundation, Inc., 59 Temple Place - Suite 330,
17 ** Boston, MA  02111-1307, USA.
18 **
19 ** Author contact information:
20 **   [email protected]
21 **   http://www.hwaci.com/drh/
22 **
23 *************************************************************************
24 ** An tokenizer for SQL
25 **
26 ** This file contains C code that splits an SQL input string up into
27 ** individual tokens and sends those tokens one-by-one over to the
28 ** parser for analysis.
29 **
30 ** $Id: tokenize.c,v 1.13 2000/08/09 17:17:25 drh Exp $
31 */
32 #include "sqliteInt.h"
33 #include <ctype.h>
34 #include <stdlib.h>
35 
36 /*
37 ** All the keywords of the SQL language are stored as in a hash
38 ** table composed of instances of the following structure.
39 */
40 typedef struct Keyword Keyword;
41 struct Keyword {
42   char *zName;             /* The keyword name */
43   int len;                 /* Number of characters in the keyword */
44   int tokenType;           /* The token value for this keyword */
45   Keyword *pNext;          /* Next keyword with the same hash */
46 };
47 
48 /*
49 ** These are the keywords
50 */
51 static Keyword aKeywordTable[] = {
52   { "ALL",               0, TK_ALL,              0 },
53   { "AND",               0, TK_AND,              0 },
54   { "AS",                0, TK_AS,               0 },
55   { "ASC",               0, TK_ASC,              0 },
56   { "BETWEEN",           0, TK_BETWEEN,          0 },
57   { "BY",                0, TK_BY,               0 },
58   { "CHECK",             0, TK_CHECK,            0 },
59   { "CONSTRAINT",        0, TK_CONSTRAINT,       0 },
60   { "COPY",              0, TK_COPY,             0 },
61   { "CREATE",            0, TK_CREATE,           0 },
62   { "DEFAULT",           0, TK_DEFAULT,          0 },
63   { "DELETE",            0, TK_DELETE,           0 },
64   { "DELIMITERS",        0, TK_DELIMITERS,       0 },
65   { "DESC",              0, TK_DESC,             0 },
66   { "DISTINCT",          0, TK_DISTINCT,         0 },
67   { "DROP",              0, TK_DROP,             0 },
68   { "EXCEPT",            0, TK_EXCEPT,           0 },
69   { "EXPLAIN",           0, TK_EXPLAIN,          0 },
70   { "FROM",              0, TK_FROM,             0 },
71   { "GLOB",              0, TK_GLOB,             0 },
72   { "GROUP",             0, TK_GROUP,            0 },
73   { "HAVING",            0, TK_HAVING,           0 },
74   { "IN",                0, TK_IN,               0 },
75   { "INDEX",             0, TK_INDEX,            0 },
76   { "INSERT",            0, TK_INSERT,           0 },
77   { "INTERSECT",         0, TK_INTERSECT,        0 },
78   { "INTO",              0, TK_INTO,             0 },
79   { "IS",                0, TK_IS,               0 },
80   { "ISNULL",            0, TK_ISNULL,           0 },
81   { "KEY",               0, TK_KEY,              0 },
82   { "LIKE",              0, TK_LIKE,             0 },
83   { "NOT",               0, TK_NOT,              0 },
84   { "NOTNULL",           0, TK_NOTNULL,          0 },
85   { "NULL",              0, TK_NULL,             0 },
86   { "ON",                0, TK_ON,               0 },
87   { "OR",                0, TK_OR,               0 },
88   { "ORDER",             0, TK_ORDER,            0 },
89   { "PRIMARY",           0, TK_PRIMARY,          0 },
90   { "SELECT",            0, TK_SELECT,           0 },
91   { "SET",               0, TK_SET,              0 },
92   { "TABLE",             0, TK_TABLE,            0 },
93   { "UNION",             0, TK_UNION,            0 },
94   { "UNIQUE",            0, TK_UNIQUE,           0 },
95   { "UPDATE",            0, TK_UPDATE,           0 },
96   { "USING",             0, TK_USING,            0 },
97   { "VACUUM",            0, TK_VACUUM,           0 },
98   { "VALUES",            0, TK_VALUES,           0 },
99   { "WHERE",             0, TK_WHERE,            0 },
100 };
101 
102 /*
103 ** This is the hash table
104 */
105 #define KEY_HASH_SIZE 37
106 static Keyword *apHashTable[KEY_HASH_SIZE];
107 
108 
109 /*
110 ** This function looks up an identifier to determine if it is a
111 ** keyword.  If it is a keyword, the token code of that keyword is
112 ** returned.  If the input is not a keyword, TK_ID is returned.
113 */
114 static int sqliteKeywordCode(const char *z, int n){
115   int h;
116   Keyword *p;
117   if( aKeywordTable[0].len==0 ){
118     /* Initialize the keyword hash table */
119     int i;
120     int n;
121     n = sizeof(aKeywordTable)/sizeof(aKeywordTable[0]);
122     for(i=0; i<n; i++){
123       aKeywordTable[i].len = strlen(aKeywordTable[i].zName);
124       h = sqliteHashNoCase(aKeywordTable[i].zName, aKeywordTable[i].len);
125       h %= KEY_HASH_SIZE;
126       aKeywordTable[i].pNext = apHashTable[h];
127       apHashTable[h] = &aKeywordTable[i];
128     }
129   }
130   h = sqliteHashNoCase(z, n) % KEY_HASH_SIZE;
131   for(p=apHashTable[h]; p; p=p->pNext){
132     if( p->len==n && sqliteStrNICmp(p->zName, z, n)==0 ){
133       return p->tokenType;
134     }
135   }
136   return TK_ID;
137 }
138 
139 /*
140 ** Return the length of the token that begins at z[0].  Return
141 ** -1 if the token is (or might be) incomplete.  Store the token
142 ** type in *tokenType before returning.
143 */
144 int sqliteGetToken(const char *z, int *tokenType){
145   int i;
146   switch( *z ){
147     case ' ': case '\t': case '\n': case '\f': case '\r': {
148       for(i=1; z[i] && isspace(z[i]); i++){}
149       *tokenType = TK_SPACE;
150       return i;
151     }
152     case '-': {
153       if( z[1]==0 ) return -1;
154       if( z[1]=='-' ){
155         for(i=2; z[i] && z[i]!='\n'; i++){}
156         *tokenType = TK_COMMENT;
157         return i;
158       }
159       *tokenType = TK_MINUS;
160       return 1;
161     }
162     case '(': {
163       *tokenType = TK_LP;
164       return 1;
165     }
166     case ')': {
167       *tokenType = TK_RP;
168       return 1;
169     }
170     case ';': {
171       *tokenType = TK_SEMI;
172       return 1;
173     }
174     case '+': {
175       *tokenType = TK_PLUS;
176       return 1;
177     }
178     case '*': {
179       *tokenType = TK_STAR;
180       return 1;
181     }
182     case '/': {
183       *tokenType = TK_SLASH;
184       return 1;
185     }
186     case '=': {
187       *tokenType = TK_EQ;
188       return 1 + (z[1]=='=');
189     }
190     case '<': {
191       if( z[1]=='=' ){
192         *tokenType = TK_LE;
193         return 2;
194       }else if( z[1]=='>' ){
195         *tokenType = TK_NE;
196         return 2;
197       }else{
198         *tokenType = TK_LT;
199         return 1;
200       }
201     }
202     case '>': {
203       if( z[1]=='=' ){
204         *tokenType = TK_GE;
205         return 2;
206       }else{
207         *tokenType = TK_GT;
208         return 1;
209       }
210     }
211     case '!': {
212       if( z[1]!='=' ){
213         *tokenType = TK_ILLEGAL;
214         return 2;
215       }else{
216         *tokenType = TK_NE;
217         return 2;
218       }
219     }
220     case '|': {
221       if( z[1]!='|' ){
222         *tokenType = TK_ILLEGAL;
223         return 1;
224       }else{
225         *tokenType = TK_CONCAT;
226         return 2;
227       }
228     }
229     case ',': {
230       *tokenType = TK_COMMA;
231       return 1;
232     }
233     case '\'': case '"': {
234       int delim = z[0];
235       for(i=1; z[i]; i++){
236         if( z[i]==delim ){
237           if( z[i+1]==delim ){
238             i++;
239           }else{
240             break;
241           }
242         }
243       }
244       if( z[i] ) i++;
245       *tokenType = TK_STRING;
246       return i;
247     }
248     case '.': {
249       if( !isdigit(z[1]) ){
250         *tokenType = TK_DOT;
251         return 1;
252       }
253       /* Fall thru into the next case */
254     }
255     case '0': case '1': case '2': case '3': case '4':
256     case '5': case '6': case '7': case '8': case '9': {
257       *tokenType = TK_INTEGER;
258       for(i=1; z[i] && isdigit(z[i]); i++){}
259       if( z[i]=='.' ){
260         i++;
261         while( z[i] && isdigit(z[i]) ){ i++; }
262         *tokenType = TK_FLOAT;
263       }
264       if( (z[i]=='e' || z[i]=='E') &&
265            ( isdigit(z[i+1])
266             || ((z[i+1]=='+' || z[i+1]=='-') && isdigit(z[i+2]))
267            )
268       ){
269         i += 2;
270         while( z[i] && isdigit(z[i]) ){ i++; }
271         *tokenType = TK_FLOAT;
272       }else if( z[0]=='.' ){
273         *tokenType = TK_FLOAT;
274       }
275       return i;
276     }
277     case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
278     case 'g': case 'h': case 'i': case 'j': case 'k': case 'l':
279     case 'm': case 'n': case 'o': case 'p': case 'q': case 'r':
280     case 's': case 't': case 'u': case 'v': case 'w': case 'x':
281     case 'y': case 'z': case '_':
282     case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
283     case 'G': case 'H': case 'I': case 'J': case 'K': case 'L':
284     case 'M': case 'N': case 'O': case 'P': case 'Q': case 'R':
285     case 'S': case 'T': case 'U': case 'V': case 'W': case 'X':
286     case 'Y': case 'Z': {
287       for(i=1; z[i] && (isalnum(z[i]) || z[i]=='_'); i++){}
288       *tokenType = sqliteKeywordCode(z, i);
289       return i;
290     }
291     default: {
292       break;
293     }
294   }
295   *tokenType = TK_ILLEGAL;
296   return 1;
297 }
298 
299 /*
300 ** Run the parser on the given SQL string.  The parser structure is
301 ** passed in.  Return the number of errors.
302 */
303 int sqliteRunParser(Parse *pParse, char *zSql, char **pzErrMsg){
304   int nErr = 0;
305   int i;
306   void *pEngine;
307   int once = 1;
308   static FILE *trace = 0;
309   extern void *sqliteParserAlloc(void*(*)(int));
310   extern void sqliteParserFree(void*, void(*)(void*));
311   extern int sqliteParser(void*, int, ...);
312   extern void sqliteParserTrace(FILE*, char *);
313 
314   i = 0;
315   sqliteParseInfoReset(pParse);
316   pEngine = sqliteParserAlloc((void*(*)(int))malloc);
317   if( pEngine==0 ){
318     sqliteSetString(pzErrMsg, "out of memory", 0);
319     return 1;
320   }
321   sqliteParserTrace(trace, "parser: ");
322   while( nErr==0 && i>=0 && zSql[i]!=0 ){
323     int tokenType;
324 
325     pParse->sLastToken.z = &zSql[i];
326     pParse->sLastToken.n = sqliteGetToken(&zSql[i], &tokenType);
327     i += pParse->sLastToken.n;
328     if( once ){
329       pParse->sFirstToken = pParse->sLastToken;
330       once = 0;
331     }
332     switch( tokenType ){
333       case TK_SPACE:
334         break;
335       case TK_COMMENT: {
336         /* Various debugging modes can be turned on and off using
337         ** special SQL comments.  Check for the special comments
338         ** here and take approriate action if found.
339         */
340 #ifndef NDEBUG
341         char *z = pParse->sLastToken.z;
342         if( sqliteStrNICmp(z,"--parser-trace-on--",19)==0 ){
343           trace = stderr;
344           sqliteParserTrace(trace, "parser: ");
345         }else if( sqliteStrNICmp(z,"--parser-trace-off--", 20)==0 ){
346           trace = 0;
347           sqliteParserTrace(trace, "parser: ");
348         }else if( sqliteStrNICmp(z,"--vdbe-trace-on--",17)==0 ){
349           pParse->db->flags |= SQLITE_VdbeTrace;
350         }else if( sqliteStrNICmp(z,"--vdbe-trace-off--", 18)==0 ){
351           pParse->db->flags &= ~SQLITE_VdbeTrace;
352 #ifdef MEMORY_DEBUG
353         }else if( sqliteStrNICmp(z,"--malloc-fail=",14)==0 ){
354           sqlite_iMallocFail = atoi(&z[14]);
355         }else if( sqliteStrNICmp(z,"--malloc-stats--", 16)==0 ){
356           if( pParse->xCallback ){
357             static char *azName[4] = {"malloc", "free", "to_fail", 0 };
358             char *azArg[4];
359             char zVal[3][30];
360             sprintf(zVal[0],"%d", sqlite_nMalloc);
361             sprintf(zVal[1],"%d", sqlite_nFree);
362             sprintf(zVal[2],"%d", sqlite_iMallocFail);
363             azArg[0] = zVal[0];
364             azArg[1] = zVal[1];
365             azArg[2] = zVal[2];
366             azArg[3] = 0;
367             pParse->xCallback(pParse->pArg, 3, azArg, azName);
368           }
369 #endif
370         }
371 #endif
372         break;
373       }
374       case TK_ILLEGAL:
375         sqliteSetNString(pzErrMsg, "unrecognized token: \"", -1,
376            pParse->sLastToken.z, pParse->sLastToken.n, "\"", 1, 0);
377         nErr++;
378         break;
379       default:
380         sqliteParser(pEngine, tokenType, pParse->sLastToken, pParse);
381         if( pParse->zErrMsg && pParse->sErrToken.z ){
382           sqliteSetNString(pzErrMsg, "near \"", -1,
383              pParse->sErrToken.z, pParse->sErrToken.n,
384              "\": ", -1,
385              pParse->zErrMsg, -1,
386              0);
387           nErr++;
388           sqliteFree(pParse->zErrMsg);
389           pParse->zErrMsg = 0;
390         }
391         break;
392     }
393   }
394   if( nErr==0 ){
395     sqliteParser(pEngine, 0, pParse->sLastToken, pParse);
396     if( pParse->zErrMsg && pParse->sErrToken.z ){
397        sqliteSetNString(pzErrMsg, "near \"", -1,
398           pParse->sErrToken.z, pParse->sErrToken.n,
399           "\": ", -1,
400           pParse->zErrMsg, -1,
401           0);
402        nErr++;
403        sqliteFree(pParse->zErrMsg);
404        pParse->zErrMsg = 0;
405     }
406   }
407   sqliteParserFree(pEngine, free);
408   if( pParse->zErrMsg ){
409     if( pzErrMsg ){
410       sqliteFree(*pzErrMsg);
411       *pzErrMsg = pParse->zErrMsg;
412     }else{
413       sqliteFree(pParse->zErrMsg);
414     }
415     if( !nErr ) nErr++;
416   }
417   if( pParse->pVdbe ){
418     sqliteVdbeDelete(pParse->pVdbe);
419     pParse->pVdbe = 0;
420   }
421   if( pParse->pNewTable ){
422     sqliteDeleteTable(pParse->db, pParse->pNewTable);
423     pParse->pNewTable = 0;
424   }
425   sqliteParseInfoReset(pParse);
426   return nErr;
427 }
428