Index: ext/fts5/fts5.h ================================================================== --- ext/fts5/fts5.h +++ ext/fts5/fts5.h @@ -107,10 +107,16 @@ ** xInst: ** Query for the details of phrase match iIdx within the current row. ** Phrase matches are numbered starting from zero, so the iIdx argument ** should be greater than or equal to zero and smaller than the value ** output by xInstCount(). +** +** Usually, output parameter *piPhrase is set to the phrase number, *piCol +** to the column in which it occurs and *piOff the token offset of the +** first token of the phrase. The exception is if the table was created +** with the offsets=0 option specified. In this case *piOff is always +** set to -1. ** ** Returns SQLITE_OK if successful, or an error code (i.e. SQLITE_NOMEM) ** if an error occurs. ** ** xRowid: @@ -194,11 +200,11 @@ ** through instances of phrase iPhrase, use the following code: ** ** Fts5PhraseIter iter; ** int iCol, iOff; ** for(pApi->xPhraseFirst(pFts, iPhrase, &iter, &iCol, &iOff); -** iOff>=0; +** iCol>=0; ** pApi->xPhraseNext(pFts, &iter, &iCol, &iOff) ** ){ ** // An instance of phrase iPhrase at offset iOff of column iCol ** } ** @@ -208,11 +214,11 @@ ** ** xPhraseNext() ** See xPhraseFirst above. */ struct Fts5ExtensionApi { - int iVersion; /* Currently always set to 1 */ + int iVersion; /* Currently always set to 2 */ void *(*xUserData)(Fts5Context*); int (*xColumnCount)(Fts5Context*); int (*xRowCount)(Fts5Context*, sqlite3_int64 *pnRow); Index: ext/fts5/fts5_main.c ================================================================== --- ext/fts5/fts5_main.c +++ ext/fts5/fts5_main.c @@ -306,10 +306,17 @@ ** Return true if pTab is a contentless table. */ static int fts5IsContentless(Fts5Table *pTab){ return pTab->pConfig->eContent==FTS5_CONTENT_NONE; } + +/* +** Return true if pTab is an offsetless table. +*/ +static int fts5IsOffsetless(Fts5Table *pTab){ + return pTab->pConfig->bOffsets==0; +} /* ** Delete a virtual table handle allocated by fts5InitVtab(). */ static void fts5FreeVtab(Fts5Table *pTab){ @@ -1747,10 +1754,14 @@ if( CsrFlagTest(pCsr, FTS5CSR_REQUIRE_INST)==0 || SQLITE_OK==(rc = fts5CacheInstArray(pCsr)) ){ if( iIdx<0 || iIdx>=pCsr->nInstCount ){ rc = SQLITE_RANGE; + }else if( fts5IsOffsetless((Fts5Table*)pCsr->base.pVtab) ){ + *piPhrase = pCsr->aInst[iIdx*3]; + *piCol = pCsr->aInst[iIdx*3 + 2]; + *piOff = -1; }else{ *piPhrase = pCsr->aInst[iIdx*3]; *piCol = pCsr->aInst[iIdx*3 + 1]; *piOff = pCsr->aInst[iIdx*3 + 2]; } @@ -1911,10 +1922,15 @@ Fts5PhraseIter *pIter, int *piCol, int *piOff ){ if( pIter->a>=pIter->b ){ *piCol = -1; + *piOff = -1; + }else if( fts5IsOffsetless((Fts5Table*)(((Fts5Cursor*)pCtx)->base.pVtab)) ){ + int iVal; + pIter->a += fts5GetVarint32(pIter->a, iVal); + *piCol += (iVal-2); *piOff = -1; }else{ int iVal; pIter->a += fts5GetVarint32(pIter->a, iVal); if( iVal==1 ){ Index: ext/fts5/fts5_tcl.c ================================================================== --- ext/fts5/fts5_tcl.c +++ ext/fts5/fts5_tcl.c @@ -233,10 +233,11 @@ { "xQueryPhrase", 2, "PHRASE SCRIPT" }, /* 11 */ { "xSetAuxdata", 1, "VALUE" }, /* 12 */ { "xGetAuxdata", 1, "CLEAR" }, /* 13 */ { "xSetAuxdataInt", 1, "INTEGER" }, /* 14 */ { "xGetAuxdataInt", 1, "CLEAR" }, /* 15 */ + { "xPhraseForeach", 4, "IPHRASE COLVAR OFFVAR SCRIPT" }, /* 16 */ { 0, 0, 0} }; int rc; int iSub = 0; @@ -426,10 +427,40 @@ int iVal; int bClear; if( Tcl_GetBooleanFromObj(interp, objv[2], &bClear) ) return TCL_ERROR; iVal = ((char*)p->pApi->xGetAuxdata(p->pFts, bClear) - (char*)0); Tcl_SetObjResult(interp, Tcl_NewIntObj(iVal)); + break; + } + + CASE(16, "xPhraseForeach") { + int iPhrase; + int iCol; + int iOff; + const char *zColvar; + const char *zOffvar; + Tcl_Obj *pScript = objv[5]; + Fts5PhraseIter iter; + + if( Tcl_GetIntFromObj(interp, objv[2], &iPhrase) ) return TCL_ERROR; + zColvar = Tcl_GetString(objv[3]); + zOffvar = Tcl_GetString(objv[4]); + + for(p->pApi->xPhraseFirst(p->pFts, iPhrase, &iter, &iCol, &iOff); + iCol>=0; + p->pApi->xPhraseNext(p->pFts, &iter, &iCol, &iOff) + ){ + Tcl_SetVar2Ex(interp, zColvar, 0, Tcl_NewIntObj(iCol), 0); + Tcl_SetVar2Ex(interp, zOffvar, 0, Tcl_NewIntObj(iOff), 0); + rc = Tcl_EvalObjEx(interp, pScript, 0); + if( rc==TCL_CONTINUE ) rc = TCL_OK; + if( rc!=TCL_OK ){ + if( rc==TCL_BREAK ) rc = TCL_OK; + break; + } + } + break; } default: assert( 0 ); Index: ext/fts5/fts5_test_mi.c ================================================================== --- ext/fts5/fts5_test_mi.c +++ ext/fts5/fts5_test_mi.c @@ -132,11 +132,11 @@ int iCol, iOff; u32 *aOut = (u32*)pUserData; int iPrev = -1; for(pApi->xPhraseFirst(pFts, 0, &iter, &iCol, &iOff); - iOff>=0; + iCol>=0; pApi->xPhraseNext(pFts, &iter, &iCol, &iOff) ){ aOut[iCol*3+1]++; if( iCol!=iPrev ) aOut[iCol*3 + 2]++; iPrev = iCol; Index: ext/fts5/test/fts5_common.tcl ================================================================== --- ext/fts5/test/fts5_common.tcl +++ ext/fts5/test/fts5_common.tcl @@ -25,10 +25,22 @@ for {set i 0} {$i < [$cmd xInstCount]} {incr i} { lappend res [string map {{ } .} [$cmd xInst $i]] } set res } + +proc fts5_test_poslist2 {cmd} { + set res [list] + + for {set i 0} {$i < [$cmd xPhraseCount]} {incr i} { + $cmd xPhraseForeach $i c o { + lappend res $i.$c.$o + } + } + + set res +} proc fts5_test_columnsize {cmd} { set res [list] for {set i 0} {$i < [$cmd xColumnCount]} {incr i} { lappend res [$cmd xColumnSize $i] @@ -111,10 +123,11 @@ foreach f { fts5_test_columnsize fts5_test_columntext fts5_test_columntotalsize fts5_test_poslist + fts5_test_poslist2 fts5_test_tokenize fts5_test_rowcount fts5_test_all fts5_test_queryphrase Index: ext/fts5/test/fts5offsets.test ================================================================== --- ext/fts5/test/fts5offsets.test +++ ext/fts5/test/fts5offsets.test @@ -72,12 +72,95 @@ INSERT INTO t2(a) VALUES('aa ab'); } #db eval {SELECT rowid, fts5_decode(rowid, block) aS r FROM t2_data} {puts $r} -breakpoint do_execsql_test 2.1 { INSERT INTO t2(t2) VALUES('integrity-check'); +} + +#------------------------------------------------------------------------- +# Check that the xInstCount, xInst, xPhraseFirst and xPhraseNext APIs +# work with offsets=0 tables. +# +set data { + 1 {abb aca aca} {aba bab aab aac caa} {abc cbc ccb bcc bab ccb aca} + 2 {bca aca acb} {ccb bcc bca aab bcc} {bab aaa aac cbb bba aca abc} + 3 {cca abc cab} {aab aba bcc cac baa} {bab cbb acb aba aab ccc cca} + 4 {ccb bcb aba} {aba bbb bcc cac bbb} {cbb aaa bca bcc aab cac aca} + 5 {bca bbc cac} {aba cbb cac cca aca} {cab acb cbc ccb cac bbb bcb} + 6 {acc bba cba} {bab bbc bbb bcb aca} {bca ccc cbb aca bac ccc ccb} + 7 {aba bab aaa} {abb bca aac bcb bcc} {bcb bbc aba aaa cba abc acc} + 8 {cab aba aaa} {ccb aca caa bbc bcc} {aaa abc ccb bbb cac cca abb} + 9 {bcb bab bac} {bcb cba cac bbb abc} {aba aca cbb acb abb ccc ccb} + 10 {aba aab ccc} {abc ccc bcc cab bbb} {aab bcc cbb ccc aaa bac baa} + 11 {bab acb cba} {aac cab cab bca cbc} {aab cbc aac baa ccb acc cac} + 12 {ccc cbb cbc} {aaa aab bcc aac bbc} {cbc cbc bac bac ccc bbc acc} + 13 {cab bbc abc} {bbb bab bba aca bab} {baa bbb aab bbb ccb bbb ccc} + 14 {bbc cab caa} {acb aac abb cba acc} {cba bba bba acb abc abb baa} + 15 {aba cca bcc} {aaa acb abc aab ccb} {cca bcb acc aaa caa cca cbc} + 16 {bcb bba aba} {cbc acb cab caa ccb} {aac aaa bbc cab cca cba abc} + 17 {caa cbb acc} {ccb bcb bca aaa bcc} {bbb aca bcb bca cbc cbc cca} + 18 {cbb bbc aac} {ccc bbc aaa aab baa} {cab cab cac cca bbc abc bbc} + 19 {ccc acc aaa} {aab cbb bca cca caa} {bcb aca aca cab acc bac bcc} + 20 {aab ccc bcb} {bbc cbb bbc aaa bcc} {cbc aab ccc aaa bcb bac cbc} + 21 {aba cab ccc} {bbc cbc cba acc bbb} {acc aab aac acb aca bca acb} + 22 {bcb bca baa} {cca bbc aca ccb cbb} {aab abc bbc aaa cab bcc bcc} + 23 {cac cbb caa} {bbc aba bbb bcc ccb} {bbc bbb cab bbc cac abb acc} + 24 {ccb acb caa} {cab bba cac bbc aac} {aac bca abc cab bca cab bcb} + 25 {bbb aca bca} {bcb acc ccc cac aca} {ccc acb acc cac cac bba bbc} + 26 {bab acc caa} {caa cab cac bac aca} {aba cac caa acc bac ccc aaa} + 27 {bca bca aaa} {ccb aca bca aaa baa} {bab acc aaa cca cba cca bac} + 28 {ccb cac cac} {bca abb bba bbc baa} {aca ccb aac cab ccc cab caa} + 29 {abc bca cab} {cac cbc cbb ccc bcc} {bcc aaa aaa acc aac cac aac} + 30 {aca acc acb} {aab aac cbb caa acb} {acb bbc bbc acc cbb bbc aac} + 31 {aba aca baa} {aca bcc cab bab acb} {bcc acb baa bcb bbc acc aba} + 32 {abb cbc caa} {cba abb bbb cbb aca} {bac aca caa cac caa ccb bbc} + 33 {bcc bcb bcb} {cca cab cbc abb bab} {caa bbc aac bbb cab cba aaa} + 34 {caa cab acc} {ccc ccc bcc acb bcc} {bac bba aca bcb bba bcb cac} + 35 {bac bcb cba} {bcc acb bbc cba bab} {abb cbb abc abc bac acc cbb} + 36 {cab bab ccb} {bca bba bab cca acc} {acc aab bcc bac acb cbb caa} + 37 {aca cbc cab} {bba aac aca aac aaa} {baa cbb cba aba cab bca bcb} + 38 {acb aab baa} {baa bab bca bbc bbb} {abc baa acc aba cab baa cac} + 39 {bcb aac cba} {bcb baa caa cac bbc} {cbc ccc bab ccb bbb caa aba} + 40 {cba ccb abc} {cbb caa cba aac bab} {cbb bbb bca bbb bac cac bca} +} +foreach {tn tbl} { + 1 { CREATE VIRTUAL TABLE t3 USING fts5(x, y, z, offsets=0) } +} { + reset_db + fts5_aux_test_functions db + execsql $tbl + foreach {id x y z} $data { + execsql { INSERT INTO t3(rowid, x, y, z) VALUES($id, $x, $y, $z) } + } + foreach {tn2 expr} { + 1 aaa 2 ccc 3 bab 4 aac + 5 aa* 6 cc* 7 ba* 8 aa* + 9 a* 10 b* 11 c* + } { + + set res [list] + foreach {id x y z} $data { + if {[lsearch [concat $x $y $z] $expr]>=0} { + lappend res $id + set inst [list] + if {[lsearch $x $expr]>=0} { lappend inst 0.0.-1 } + if {[lsearch $y $expr]>=0} { lappend inst 0.1.-1 } + if {[lsearch $z $expr]>=0} { lappend inst 0.2.-1 } + lappend res $inst + } + } + + do_execsql_test 3.$tn.$tn2.1 { + SELECT rowid, fts5_test_poslist(t3) FROM t3($expr) + } $res + + do_execsql_test 3.$tn.$tn2.2 { + SELECT rowid, fts5_test_poslist2(t3) FROM t3($expr) + } $res + } + } finish_test