>>> BTW, how many companies do you know that use the CF Verity engine
>>> successfully?
>> I know plenty, myself. That's not to say it's perfect (or even good), but it
>> is adequate for many, at least. It also has the advantage of reliably being
>> there for most CF installations, which can potentially reduce the
>> installation dependencies of CF applications.
>
> I remember having to index a rather small table with only 100,000 rows, and
> by the end of the day the search would just crawl. I also didn't appreciate
> having to schedule a bunch of jobs in order to optimize the collections.
This would imply that the code was looping over the recordset and
updating the Verity collection one record at a time. Inevitably this
leads to massive fragmentation of the Verity index and consequentially
VERY slow performance until you optimise the collection.
Obviously there are indexing approaches that would minimise the
fragmentation of the index in the first place. eg. updating the
collection with the entire query object.
As for optimising, you can set up a single job to automatically optimise
all the collections on a single server:
Verity Optimisation Made Easy
http://blog.daemon.com.au/archives/000073.html
-- geoff
http://www.daemon.com.au/
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
