Since we can have multiple FTS indexes on a table, we keep a per index cache of words etc. More...

#include <fts0types.h>

Public Attributes
dict_index_t *	index
	The FTS index instance. More...

ib_rbt_t *	words
	Nodes; indexed by fts_string_t, cells are fts_tokenizer_word_t. More...

ib_vector_t *	doc_stats
	Array of the fts_doc_stats_t contained in the memory buffer. More...

que_t **	ins_graph
	Insert query graphs. More...

que_t **	sel_graph
	Select query graphs. More...

CHARSET_INFO *	charset
	charset More...

Detailed Description

Since we can have multiple FTS indexes on a table, we keep a per index cache of words etc.

Member Data Documentation

CHARSET_INFO* fts_index_cache_t::charset

charset

ib_vector_t* fts_index_cache_t::doc_stats

Array of the fts_doc_stats_t contained in the memory buffer.

Must be in sorted order (ascending). The ideal choice is an rb tree but the rb tree imposes a space overhead that we can do without

dict_index_t* fts_index_cache_t::index

The FTS index instance.

que_t** fts_index_cache_t::ins_graph

Insert query graphs.

que_t** fts_index_cache_t::sel_graph

Select query graphs.

ib_rbt_t* fts_index_cache_t::words

Nodes; indexed by fts_string_t*, cells are fts_tokenizer_word_t*.

The documentation for this struct was generated from the following file: