[Xapian-discuss] re-indexing performance using omindex

tigerf tigerf at web.de
Wed May 20 11:22:18 BST 2009


Hi list,

I'm using xapian/omega V1.0.7 to index 50.000+ mainly .doc and .pdf
files residing in a directory subtree on a virtualized Ubuntu 8.10
server. Querying/presenting is done via PHP, what works remarkably well
and fast, compliments.

Allmost all works as expected. Initial indexing takes 2.5 hrs, what is
acceptable. But when re-indexing the directory it still takes the same
time to finish, even if no or only few files were added/changed meanwhile.

Apparently omindex re-reads all files every time, not only those that
are new, have timestamps or size changed.

Is there an option to make omindex updating only the
modified/new/deleted files, leaving the rest untouched?
I would expect this as the default behaviour.
"-p" didn't help.

May be a newbie question, but I didn't find a solution yet.

Thx
Tiger




More information about the Xapian-discuss mailing list