>Hi!
>
>You can download it through HTTP using wwIPStuff library from West-Wind programmatically to disk, then load into the MSXML parser from disk file.
>
>Personally, I always beleived that 200MB XML file is not a case in a real life, 1-2 MB maximum.
>XML parser probably not able to load such large file, or it could take so much time, that the file transfer over the internet is disconnected because timout ;).
>
>Just a suggestion. I was not able to find neither system capacities of the MSXML parser, nor tests or samples for such large data file.

I don't think there are any system capacities beyond the hardware you're running. I've worked with some fairly large files (5-10megs) and it was acceptable in terms of speed, but you could tell this is pushing hte limits of where the in memory footprint of the document was overwhelming the parser logic to find nodes and walk the tree.

As to downloads, XMLHTTP is all memory based and doing a 200 meg download is not going to work unless you have a boatload of memory.

You can use WinInet to do this (or wwIPStuff which wraps it) and allows you to write the data to disk. In wwIPStuff you can specify a file to dump incoming HTTP data to as the data gets downloaded to minimize memory usage on the client for large downloads.

Timeouts are determined by the Web Server not the client so setting up the server to not throw a timeout message is key to doing large downloads. I can't remember now offhand, but I think straight file downloads don't time out unless the data stream stops.