I'm not an expert on this, but at midnight you get who's online. :) Let's assume it takes 4-5 seconds to process (i.e. cache) a newly modified dictionary.
Let's also assume MarkLogic is lazy about processing new dictionaries until they're first used. What you were doing before was "priming the pump" by using the dictionary in the spawn so it'd be processed before the first real user needed it. Now you're saying it's still slow on the first request even after the spawn() completes? That seems unlikely. -jh- On Apr 9, 2012, at 11:10 PM, Yeshmeet Singh wrote: > Hi, > > We are using spell:suggest and the spell:suggest-detailed api's for fetching > the correct words from a custom spell dictionary. Now if we update the spell > dictionary and then fire the query, it takes a long time to return the > results for the first hit. Again if we fire the same query the time taken to > return back the results is considerably reduced. To fix the issue I > asynchronously called my xquery file using Xdmp:Spawn() after the insertion. > It was working fine for me till I used a limited number of dictionaries in my > spawn call. But now I had to include some more dictionaries and it is again > taking around 4-5 seconds to fetch after a new word is inserted in the > dictionary. Is there any solution to the problem? > > Thanks > Yashmeet Singh
_______________________________________________ General mailing list [email protected] http://developer.marklogic.com/mailman/listinfo/general
