Does anyone know if there is a limit to how large a document can be before webMethods cannot handle it? XML files have a tendancy to get quite large, especially when using the Rosettanet Standard. I am running on a Unix server that contains 1 GB of memory solely for Java. Using this server, it is able to process an XML file in excess of 25 MB, but it takes up the entire usage of the CPU. We are afraid this may cause problems if any other webMethods process tries to run as well. Smaller XML files do not have this problem. Is there any documentation on how large a file can be or something similar that may help. Thanks for any responses.

• John Oggenfuss

The XML parser in the Integration Server supports streaming-parse-on-demand, therefore, there the complete XML file never has to be in memory to be parsed.

That being said, it depends what operations you are performing on the XML file. There is a document or section of a document on Large Document Handling in most of the Integration Platform Components (Trading Networks, EDI, etc).

For instance, XQL queries require the complete document to be in memory. pub.web:documentToRecord creates an IData representation of the complete document, but pub.web:nodeIterator supports a “moving window” of parsed XML, which can support arbitrarily large documents.