On Tue, 2003-12-09 at 12:24, Rajesh Raheja wrote: > Thanks for the replies. > > I will lookup "pruning" and "filtering" on this mailing list and also research > STX. Although I would appreciate more insight on the SAX filtering usage > pattern for large documents and how I would be able to apply it.
Joe's mention of "pruning" and "filtering" is referring to features that may be implemented in Xalan sometime in the next few years. Unfortunately it won't help you now. Re the "SAX filtering": suppose your input document contains large numbers of address tags, and your desired output doesn't contain any address info, then you can optimise processing by discarding all the address tags & their children before the XSLT engine sees the input. See the SAX class org.xml.sax.XMLFilter for more info. However if your mapping actually uses most/all of the input data then this won't help. If your input data is composed of multiple independent "pieces", then you have another option: split the input data into multiple smaller documents and apply the mapping to each piece at a time. This can also give a nice performance boost. However it obviously only applies for some types of input data. Regards, Simon
