In general, Corb is an aggressively minimalist tool. It does content reprocessing in bulk, and does so with a bare minimum of options.
Corb does support arbitrary queries, with the URIS-MODULE option. Simply write the query that returns the uris you want to return. For directories, use cts:uris() with a cts:directory-query(). https://github.com/marklogic/corb/blob/master/src/java/com/marklogic/developer/corb/get-uris.xqy has an example that you could adapt to your needs. For thread count, I would start with two threads per server CPU core. If that isn't fast enough, try doubling it and see if the operation rate goes up. If it doesn't, then look at optimizing your XQuery module - or possibly tuning the server. -- Mike On 13 Jul 2011, at 04:46 , sachin gill wrote: > Hi, > I have just implemented CORB (implemented as jar)as a processing tool ! > I'm successfully able to run this tool.But I found there are some > limitations to it: > 1) It reads only collection type and not other types(like directory , > Query) etc. 5) If somebody can explain what all ways I can use corb it will > be great. > 2) Need to know how many optimum threads should I run to optimize it > running time. > 3) Currently I've a restriction that my getURI.xqy file and Module URI > need to be under same path, because if add extra path to module uri or > processing file it doesn't work. I want to use different Module URI's files. > 4)I'll really appreciate If I get an explanation for how broad i can > use this tool , it will be great because i want it to be used for all sort of > bulk processing like appending XML elements, svn update. If so then how. > > Thanks > Sachin > > _______________________________________________ > General mailing list > General@developer.marklogic.com > http://developer.marklogic.com/mailman/listinfo/general _______________________________________________ General mailing list General@developer.marklogic.com http://developer.marklogic.com/mailman/listinfo/general
