boolean fulltext search without an index (4a5ca107) · Commits · Software / OSDI20 Artifacts / mariadb

Docs/manual.texi

+10 −6

Original line number	Diff line number	Diff line
		@@ -29167,8 +29167,6 @@ mysql> select STRCMP('text', 'text');
		relevance - similarity measure between the text in columns
		@code{(col1,col2,...)} and the query @code{expr}. Relevance is a
		positive floating-point number. Zero relevance means no similarity.
		For @code{MATCH ... AGAINST()} to work, a @strong{FULLTEXT} index
		must be created first. @xref{CREATE TABLE, , @code{CREATE TABLE}}.
		@code{MATCH ... AGAINST()} is available in MySQL version
		3.23.23 or later. @code{IN BOOLEAN MODE} extension was added in version
		4.0.1. For details and usage examples @pxref{Fulltext Search}.
		@@ -33828,9 +33826,10 @@ mysql> SELECT * FROM articles WHERE MATCH (title,body) AGAINST (

		This query retrieved all the rows that contain the word @code{MySQL}
		(note: 50% threshold is gone), but does @strong{not} contain the word
		@code{YourSQL}. Note that it does not auto-magically sort rows in
		@code{YourSQL}. Note, that it does not auto-magically sort rows in
		derceasing relevance order (the last row has the highest relevance,
		as it contains @code{MySQL} twice).
		as it contains @code{MySQL} twice). Boolean fulltext search can also
		work even without @code{FULLTEXT} index, but it would be @strong{slow}.

		Boolean fulltext search supports the following operators:

		@@ -33890,10 +33889,12 @@ order), but rank ``gates to hell'' higher than ``bill gates''.
		@itemize @bullet
		@item
		All parameters to the @code{MATCH} function must be columns from the
		same table that is part of the same fulltext index.
		same table that is part of the same fulltext index, unless this
		@code{MATCH} is @code{IN BOOLEAN MODE}.
		@item
		Column list between @code{MATCH} and @code{AGAINST} must match exactly
		a column list in the @code{FULLTEXT} index definition.
		a column list in the @code{FULLTEXT} index definition, unless this
		@code{MATCH} is @code{IN BOOLEAN MODE}.
		@item
		The argument to @code{AGAINST} must be a constant string.
		@end itemize
		@@ -45853,6 +45854,9 @@ Our TODO section contains what we plan to have in 4.0. @xref{TODO MySQL 4.0}.

		@itemize @bullet
		@item
		@code{MATCH ... AGAINST(... IN BOOLEAN MODE)} can now work
		without @code{FULLTEXT} index.
		@item
		Added @file{myisam/ft_dump} utility for low-level inspection
		of @code{FULLTEXT} indexes.
		@item

include/ft_global.h

+1 −1

Original line number	Diff line number	Diff line
		@@ -32,7 +32,7 @@ extern "C" {
		typedef struct st_ft_info FT_INFO;
		struct _ft_vft {
		int (read_next)(FT_INFO , char *);
		float (find_relevance)(FT_INFO , my_off_t, byte *);
		float (find_relevance)(FT_INFO , byte *, uint);
		void (close_search)(FT_INFO );
		float (get_relevance)(FT_INFO );
		void (reinit_search)(FT_INFO );

include/my_base.h

+1 −0

Original line number	Diff line number	Diff line
		@@ -226,6 +226,7 @@ enum ha_base_keytype {
		/* Other constants */

		#define HA_NAMELEN 64 /* Max length of saved filename */
		#define NO_SUCH_KEY ((uint)~0) /* used as a key no. */

		/* Intern constants in databases */

myisam/ft_boolean_search.c

+15 −7

Original line number	Diff line number	Diff line
		@@ -152,13 +152,16 @@ void _ftb_init_index_search(FT_INFO *ftb)
		int i, r;
		FTB_WORD *ftbw;
		MI_INFO *info=ftb->info;
		MI_KEYDEF *keyinfo=info->s->keyinfo+ftb->keynr;
		my_off_t keyroot=info->s->state.key_root[ftb->keynr];
		MI_KEYDEF *keyinfo;
		my_off_t keyroot;

		if (ftb->state != READY)
		if (ftb->state != READY \|\| ftb->keynr == NO_SUCH_KEY)
		return;
		ftb->state=INDEX_SEARCH;

		keyinfo=info->s->keyinfo+ftb->keynr;
		keyroot=info->s->state.key_root[ftb->keynr];

		for (i=ftb->queue.elements; i; i--)
		{
		ftbw=(FTB_WORD *)(ftb->queue.root[i]);
		@@ -352,14 +355,17 @@ int ft_boolean_read_next(FT_INFO ftb, char record)
		return my_errno=HA_ERR_END_OF_FILE;
		}

		float ft_boolean_find_relevance(FT_INFO ftb, my_off_t docid, byte record)
		float ft_boolean_find_relevance(FT_INFO ftb, byte record, uint length)
		{
		TREE ptree;
		FT_WORD word;
		FTB_WORD *ftbw;
		FTB_EXPR *ftbe;
		uint i;
		my_off_t docid=ftb->info->lastpos;

		if (docid == HA_POS_ERROR)
		return -2.0;
		if (ftb->state == READY \|\| ftb->state == INDEX_DONE)
		{
		for (i=1; i<=ftb->queue.elements; i++)
		@@ -382,11 +388,13 @@ float ft_boolean_find_relevance(FT_INFO ftb, my_off_t docid, byte record)
		ftb->state=SCAN;
		}
		else if (ftb->state != SCAN)
		return -2.0;
		return -3.0;

		bzero(&ptree, sizeof(ptree));
		if (_mi_ft_parse(& ptree, ftb->info, ftb->keynr, record))
		return -3.0;
		if ((ftb->keynr==NO_SUCH_KEY)
		? ft_parse(& ptree, record, length)
		: _mi_ft_parse(& ptree, ftb->info, ftb->keynr, record))
		return -4.0;

		for (i=1; i<=ftb->queue.elements; i++)
		{

myisam/ft_dump.c

+10 −2

Original line number	Diff line number	Diff line
		@@ -159,7 +159,7 @@ int main(int argc,char *argv[])
		return 0;
		}

		const char *options="dscve:h";
		const char *options="dscvh";

		static void get_options(int argc, char *argv[])
		{
		@@ -184,7 +184,15 @@ static void get_options(int argc, char *argv[])

		static void usage(char *argv[])
		{
		printf("Use: %s [-%s] <table_name> <key_no>\n", *argv, options);
		printf("
		Use: %s [-%s] <table_name> <index_no>

		-d dump index (incl. data offsets and word weights)
		-s report global stats
		-c calculate per-word stats (counts and global weights)
		-v be verbose
		-h this text\n
		", *argv, options);
		exit(1);
		}