I compared speed of importing (into an empty SQLite DB) from 2 kinds of SQL dumps. Data is exactly the same in both cases, and xz-compressed size of SQL dump is near 18MB in both cases. First SQL dump has single big INSERT statement in single transaction. Second SQL dump has one INSERT statement for each row.
Schema is 1 table with these columns: INTEGER PRIMARY KEY, TEXT x2 (2 columns), INTEGER x4, TEXT x10 There is nothing else besides the data, no indexes, etc. in both cases. In both cases I am importing from compressed file, using command like this: time xzdec something.sql.xz | sqlite3 something.db Time of import from single-insert SQL dump: real 2m13.884s user 2m13.791s sys 0m1.052s Time of import from multiple-insert SQL dump: real 0m2.192s user 0m3.266s sys 0m0.347s IMO sqlite needs some optimisation for this case when there’s a huge INSERT statement, because the speed difference is enormous. _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
