Xml variable removes process-instruction

Question

I have a project that reads documents through a watch folder. The files that are dropped in are xml and contain a processing instruction with some meta data is needed in my process. I have configured a watch folder endpoint using 3 different type of input variables for document, xml, and string. When I do document I get the document object but no access to the processing instruction. When I do xml, I get the xml in a variable but all processing instructions are removed including the xml declaration. Finally, when I use a string, the variable is set with a binary representation of the file that looks like base 64. I have tested a different process reading the xml from file system as a string and I can use string functions to extract the values I need that way. I'm wondering if anybody knows what I can do to configure my watch folder enpoint so that I can access the processing directive either as xml or as a string.

Answer 1

Hi,
The <?xml version='1.0' encoding='utf-8'?>) xml declaration can occur only once and that too at the very top of the xml file before everything else. I think your best bet would be to strip the xml declaration and the doctype declaration before appending. You might want to check the API documentation of the API you are using to serialize the xml to see if it supports suppressing xml and doctype declarations.
With apache xerces, you can do
OutputFormat.setOmitXMLDeclaration(true);
where OutputFormat dictates what underlying serialized is used to serialize the document based on the specified method.
Hope this helps.
Cheers,
vidyut