We have an index that takes several seconds in updating. Since this is the first index in the queue, the others have to wait until it is their turn, leaving the end user with the impression of a very slow update.

Therefor we would like to explore more efficient ways, like multiple dedicated index engines (as required).

Is it possible to

a) have 2 or more dedicated index engines coexisting?

b) assign one index engine exclusively for certain files, while having another engine update the rest?

Regards

Wilhelm

Yup. Just have the indexer attach just the appropriate files.

The Sprezzatura Group

World Leaders in all things RevSoft

Depending on the amount of changed data etc, it might be easier to Index.Flush after the write of a screen (and other processes).

We had a user that liked getting alittle message with the flush.

But as mentioned it really does depend on the processes ..

[email protected] onmouseover=window.status=the new revelation technology .. a refreshing change;return(true)"

DSig's Radio Free Oregon

David Tod Sigafoos ~ SigSolutions

We activate oengine with the /i switch, so all tables are attached from an image.

How could we then separate index A,B and C for engine 1 and the rest for engine 2?

Wilhelm

Wilhelm,

The engine has the capability to execute a Basic+ procedure on startup, so with that in mind you could write a startup proc to just attach the tables you want to maintain and then use the set_idxSvr() function to start the indexing process.

The trick here is to work out what tables to attach - you could launch the engine with a specific user via the /un switch, and then attach based on the user name; or you could add your own command line switch to specify the table set, and detect it in your startup proc via the GetCommandLine() Windows API function.

The Sprezzatura Group

World leaders in all things RevSoft

View this thread on the Works forum...