Hello! On Thursday 18 March 2010 20:18:10 P Kishor wrote: > Alexey's issue is that for the same query, his compressed file is > slower than the non-compressed file.
The 'file_text_content' table content is compressed and fast. But the virtual table 'file_text' is slow for count(*) query. The SQL command CREATE VIRTUAL TABLE file_text USING fts3(content, meta, TOKENIZE icu en_US); produce this database schema: CREATE VIRTUAL TABLE file_text USING fts3(content, meta, TOKENIZE icu en_US); CREATE TABLE 'file_text_content'(docid INTEGER PRIMARY KEY, 'c0content' blob, 'c1meta' blob); CREATE TABLE 'file_text_segdir'(level INTEGER,idx INTEGER,start_block INTEGER,leaves_end_block INTEGER,end_block INTEGER,root BLOB,PRIMARY KEY(level, idx)); CREATE TABLE 'file_text_segments'(blockid INTEGER PRIMARY KEY, block BLOB); Virtual table interface enumerate rows by selecting full table content instead of index scan. This solution must be optimized but I don't know how to do it. Best regards, Alexey Pechnikov. http://pechnikov.tel/ _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
