On Mon, 2009-02-23 at 17:36 -0500, [email protected] wrote: [...] > We normally deal with very large XML files and avoid loading them into > memory in our application by using SAX api. However in scripts we > call xsltproc.
> Am I understanding it correctly that xsltproc does in fact load entire > XML into memory? > > And is that a part of a spec, or could it be redone by SAX parsing and > avoiding a large memory imprint. In general it's pretty hard to implement a streaming XSLT engine. there are people looking at a profile of XSLT that could be streamed, but in the meantime i'd suggest considering alternatives -- if you only need to fetch small parts of the document, but multiple times for a given large document, you could use XQuery with an indexed implementation like dbxml, or if you need to read the entire document only once, you could consider using STX, the streaming xml transformation language, although I'm not up to speed on implementations. Hope this helps at least a little. Liam -- Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/ Pictures from old books: http://fromoldbooks.org/ Ankh: irc.sorcery.net irc.gnome.org www.advogato.org _______________________________________________ xml mailing list, project page http://xmlsoft.org/ [email protected] http://mail.gnome.org/mailman/listinfo/xml
