Add a version of the LIKE operator to the icu extension. Requires optimisation. ...

FossilOrigin-Name: 3e96105c1f084a4ab4dad4de6f4759e43fc497f7

diff --git a/ext/icu/icu.c b/ext/icu/icu.c
index e58e834caa33497284e8e70b3444004622d774d8..7e9e8cfad8ff037b753383b6d7d66ee8e0255e90 100644 (file)
--- a/ext/icu/icu.c
+++ b/ext/icu/icu.c
@@ -5,14 +5,14 @@
 ** for handling unicode data) and SQLite. The integration uses 
 ** ICU to provide the following to SQLite:
 **
+**   * An implementation of the SQL regexp() function (and hence REGEXP
+**     operator) using the ICU uregex_XX() APIs.
+**
 **   * Implementations of the SQL scalar upper() and lower() 
-**     functions for case mapping, 
+**     functions for case mapping.
 **
 **   * Collation sequences
 **
-**   * Implementation of the SQL regexp() function (and hence REGEXP
-**     operator) using the ICU uregex_XX() APIs.
-**
 **   * LIKE
 */
 
@@ -47,10 +47,130 @@ static void xFree(void *p){
 }
 
 /*
-** LIKE operator.
+** Compare two UTF-8 strings for equality where the first string is
+** a "LIKE" expression. Return true (1) if they are the same and 
+** false (0) if they are different.
+*/
+static int icuLikeCompare(
+  const uint8_t *zPattern,   /* The UTF-8 LIKE pattern */
+  const uint8_t *zString,    /* The UTF-8 string to compare against */
+  const UChar32 uEsc         /* The escape character */
+){
+  static const int MATCH_ONE = (UChar32)'_';
+  static const int MATCH_ALL = (UChar32)'%';
+
+  int iPattern = 0;       /* Current byte index in zPattern */
+  int iString = 0;        /* Current byte index in zString */
+
+  int prevEscape = 0;     /* True if the previous character was uEsc */
+
+  while( zPattern[iPattern]!=0 ){
+
+    /* Read (and consume) the next character from the input pattern. */
+    UChar32 uPattern;
+    U8_NEXT_UNSAFE(zPattern, iPattern, uPattern);
+    assert(uPattern!=0);
+
+    /* There are now 4 possibilities:
+    **
+    **     1. uPattern is an unescaped match-all character "%",
+    **     2. uPattern is an unescaped match-one character "_",
+    **     3. uPattern is an unescaped escape character, or
+    **     4. uPattern is to be handled as an ordinary character
+    */
+    if( !prevEscape && uPattern==MATCH_ALL ){
+      /* Case 1. */
+      uint8_t c;
+
+      /* Skip any MATCH_ALL or MATCH_ONE characters that follow a
+      ** MATCH_ALL. For each MATCH_ONE, skip one character in the 
+      ** test string.
+      */
+      while( (c=zPattern[iPattern]) == MATCH_ALL || c == MATCH_ONE ){
+        if( c==MATCH_ONE ){
+          if( zString[iString]==0 ) return 0;
+          U8_FWD_1_UNSAFE(zString, iString);
+        }
+        iPattern++;
+      }
+
+      if( zPattern[iPattern]==0 ) return 1;
+
+      while( zString[iString] ){
+        if( icuLikeCompare(&zPattern[iPattern], &zString[iString], uEsc) ){
+          return 1;
+        }
+        U8_FWD_1_UNSAFE(zString, iString);
+      }
+      return 0;
+
+    }else if( !prevEscape && uPattern==MATCH_ONE ){
+      /* Case 2. */
+      if( zString[iString]==0 ) return 0;
+      U8_FWD_1_UNSAFE(zString, iString);
+
+    }else if( !prevEscape && uPattern==uEsc){
+      /* Case 3. */
+      prevEscape = 1;
+
+    }else{
+      /* Case 4. */
+      UChar32 uString;
+      U8_NEXT_UNSAFE(zString, iString, uString);
+      uString = u_foldCase(uString, U_FOLD_CASE_DEFAULT);
+      uPattern = u_foldCase(uPattern, U_FOLD_CASE_DEFAULT);
+      if( uString!=uPattern ){
+        return 0;
+      }
+      prevEscape = 0;
+    }
+  }
+
+  return zString[iString]==0;
+}
+
+/*
+** Implementation of the like() SQL function.  This function implements
+** the build-in LIKE operator.  The first argument to the function is the
+** pattern and the second argument is the string.  So, the SQL statements:
 **
-** http://unicode.org/reports/tr21/tr21-5.html#Caseless_Matching
+**       A LIKE B
+**
+** is implemented as like(B, A). If there is an escape character E, 
+**
+**       A LIKE B ESCAPE E
+**
+** is mapped to like(B, A, E).
 */
+static void icuLikeFunc(
+  sqlite3_context *context, 
+  int argc, 
+  sqlite3_value **argv
+){
+  const unsigned char *zA = sqlite3_value_text(argv[0]);
+  const unsigned char *zB = sqlite3_value_text(argv[1]);
+  UChar32 uEsc = 0;
+
+  if( argc==3 ){
+    /* The escape character string must consist of a single UTF-8 character.
+    ** Otherwise, return an error.
+    */
+    int nE= sqlite3_value_bytes(argv[2]);
+    const unsigned char *zE = sqlite3_value_text(argv[2]);
+    int i = 0;
+    if( zE==0 ) return;
+    U8_NEXT(zE, i, nE, uEsc);
+    if( i!=nE){
+      sqlite3_result_error(context, 
+          "ESCAPE expression must be a single character", -1);
+      return;
+    }
+  }
+
+  if( zA && zB ){
+    sqlite3_result_int(context, icuLikeCompare(zA, zB, uEsc));
+  }
+}
 
 /*
 ** This function is called when an ICU function called from within
@@ -194,6 +314,9 @@ static void icuCaseFunc16(sqlite3_context *p, int nArg, sqlite3_value **apArg){
   }
 
   zInput = sqlite3_value_text16(apArg[0]);
+  if( !zInput ){
+    return;
+  }
   nInput = sqlite3_value_bytes16(apArg[0]);
 
   nOutput = nInput * 2 + 2;
@@ -244,7 +367,7 @@ static int icuCollationColl(
     case UCOL_GREATER: return +1;
     case UCOL_EQUAL:   return 0;
   }
-  assert(!"Bad return value from ucol_strcoll()");
+  assert(!"Unexpected return value from ucol_strcoll()");
   return 0;
 }
 
@@ -288,7 +411,7 @@ static void icuLoadCollation(
   }
   assert(p);
 
-  rc = sqlite3_create_collation_x(db, zName, SQLITE_UTF16, (void *)pUCollator, 
+  rc = sqlite3_create_collation_v2(db, zName, SQLITE_UTF16, (void *)pUCollator, 
       icuCollationColl, icuCollationDel
   );
   if( rc!=SQLITE_OK ){
@@ -308,7 +431,7 @@ int sqlite3IcuInit(sqlite3 *db){
     void *pContext;                           /* sqlite3_user_data() context */
     void (*xFunc)(sqlite3_context*,int,sqlite3_value**);
   } scalars[] = {
-    {"regexp", 2, SQLITE_ANY,          0, icuRegexpFunc},
+    {"regexp",-1, SQLITE_ANY,          0, icuRegexpFunc},
 
     {"lower",  1, SQLITE_UTF16,        0, icuCaseFunc16},
     {"lower",  2, SQLITE_UTF16,        0, icuCaseFunc16},
@@ -320,6 +443,9 @@ int sqlite3IcuInit(sqlite3 *db){
     {"upper",  1, SQLITE_UTF8,  (void*)1, icuCaseFunc16},
     {"upper",  2, SQLITE_UTF8,  (void*)1, icuCaseFunc16},
 
+    {"like",   2, SQLITE_UTF8,         0, icuLikeFunc},
+    {"like",   3, SQLITE_UTF8,         0, icuLikeFunc},
+
     {"icu_load_collation",  2, SQLITE_UTF8, (void*)db, icuLoadCollation},
   };
 

diff --git a/manifest b/manifest
index 4e83c9788d3a3b4d24023ee82a2b4b8244571dd0..919a206fdebfa5129834f63fd6b92ff5d2e8a122 100644 (file)
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Change\sthe\sname\sof\screate_collation_x()\sto\screate_collation_v2().\sAlso\sadd\ssome\stests\sfor\sit.\s(CVS\s3938)
-D 2007-05-07T14:58:53
+C Add\sa\sversion\sof\sthe\sLIKE\soperator\sto\sthe\sicu\sextension.\sRequires\soptimisation.\s(CVS\s3939)
+D 2007-05-07T16:58:02
 F Makefile.in ab0f3cb6b34aa8ccec0bb57e6696fd4bd6b34a8f
 F Makefile.linux-gcc 2d8574d1ba75f129aba2019f0b959db380a90935
 F README 9c4e2d6706bdcc3efdd773ce752a8cdab4f90028
@@ -43,7 +43,7 @@ F ext/fts2/fts2_porter.c 991a45463553c7318063fe7773368a6c0f39e35d
 F ext/fts2/fts2_tokenizer.h 4c5ffe31d63622869eb6eec1503df7f6996fd1bd
 F ext/fts2/fts2_tokenizer1.c 5c979fe8815f95396beb22b627571da895a025af
 F ext/fts2/mkfts2amal.tcl 2a9ec76b0760fe7f3669dca5bc0d60728bc1c977
-F ext/icu/icu.c 509ac3d8afea8af6835edb9d96a52a80dd56c152
+F ext/icu/icu.c 6b47f5bbaf32bce03112282ecca1f54bec969e42
 F install-sh 9d4de14ab9fb0facae2f48780b874848cbf2f895
 F ltmain.sh 56abb507100ed2d4261f6dd1653dec3cf4066387
 F main.mk 09c19ae05ac9e5654d5fd866a980b21ad9df8f30
@@ -245,7 +245,7 @@ F test/fts2k.test 222d0b3bc8667753f18406aaea9906a6098ea016
 F test/fts2l.test 4c53c89ce3919003765ff4fd8d98ecf724d97dd3
 F test/fts2m.test 4b30142ead6f3ed076e880a2a464064c5ad58c51
 F test/fts2n.test a70357e72742681eaebfdbe9007b87ff3b771638
-F test/func.test 6727c7729472ae52b5acd86e802f89aa350ba50f
+F test/func.test 5e32fe07bf4113ce2923df28af78c76702f6cd92
 F test/hook.test 7e7645fd9a033f79cce8fdff151e32715e7ec50a
 F test/icu.test e6bfae7f625c88fd14df6f540fe835bdfc1e4329
 F test/in.test 369cb2aa1eab02296b4ec470732fe8c131260b1d
@@ -483,7 +483,7 @@ F www/tclsqlite.tcl bb0d1357328a42b1993d78573e587c6dcbc964b9
 F www/vdbe.tcl 87a31ace769f20d3627a64fa1fade7fed47b90d0
 F www/version3.tcl 890248cf7b70e60c383b0e84d77d5132b3ead42b
 F www/whentouse.tcl fc46eae081251c3c181bd79c5faef8195d7991a5
-P db51f59a7bb7530f919858e2c51057839f6c9f4d
-R 27185812931e7f9d41fbfd4819a22118
+P ddc4e4797ff902692c4f0d86ec5f4e94cc7f0741
+R c3baf3c645e55d9b878c7a05ae7a30bb
 U danielk1977
-Z ae7686d308deffcdeb3aacd0475a723f
+Z a53ea51c5cda49495951728b7ccf7458

diff --git a/manifest.uuid b/manifest.uuid
index fc261798e1c2faa99ce1dde1c3d1c3e7ac04fef0..9079546f3c4bdf7fa441e0ae643465b17299d532 100644 (file)
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-ddc4e4797ff902692c4f0d86ec5f4e94cc7f0741
\ No newline at end of file
+3e96105c1f084a4ab4dad4de6f4759e43fc497f7
\ No newline at end of file

diff --git a/test/func.test b/test/func.test
index c1389b3535475caa3150ac6db588a4fd822256a8..f64c8b71216aedef2483a6249f83937108478fe7 100644 (file)
--- a/test/func.test
+++ b/test/func.test
@@ -11,7 +11,7 @@
 # This file implements regression tests for SQLite library.  The
 # focus of this file is testing built-in functions.
 #
-# $Id: func.test,v 1.63 2007/05/02 15:36:02 drh Exp $
+# $Id: func.test,v 1.64 2007/05/07 16:58:02 danielk1977 Exp $
 
 set testdir [file dirname $argv0]
 source $testdir/tester.tcl
@@ -220,9 +220,11 @@ do_test func-5.2 {
 do_test func-5.3 {
   execsql {SELECT upper(a), lower(a) FROM t2}
 } {1 1 {} {} 345 345 {} {} 67890 67890}
-do_test func-5.4 {
-  catchsql {SELECT upper(a,5) FROM t2}
-} {1 {wrong number of arguments to function upper()}}
+ifcapable !icu {
+  do_test func-5.4 {
+    catchsql {SELECT upper(a,5) FROM t2}
+  } {1 {wrong number of arguments to function upper()}}
+}
 do_test func-5.5 {
   catchsql {SELECT upper(*) FROM t2}
 } {1 {wrong number of arguments to function upper()}}