On 1/06/2009 5:29 PM, pierr wrote: > Hi all, > > Section 7.9 of http://www.sqlite.org/atomiccommit.html mentioned: > > "On embedded systems with synchronous filesystems, TRUNCATE results in > slower behavior than PERSIST. The commit operation is the same speed. But > subsequent transactions are slower following a TRUNCATE because it is faster > to overwrite existing content than to append to the end of a file. New > journal file entries will always be appended following a TRUNCATE but will > usually overwrite with PERSIST." > > why " New journal file entries will always be appended following a TRUNCATE > but will usually overwrite with PERSIST"? I think if we trancate the > journey file to _zero_ , and the following write to that file is going to > overwrite the old data. Why append?
Truncate means to set the end of the file to a smaller value than before; pages so freed are then owned by the filesystem. Writing more data then requires appending which requires pages to be obtained from the filesystem, which is not a zero-cost operation. It may write over old data (you get your old pages back), or it may not (someone else has grabbed your freed pages in the meantime). HTH, John _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
