As part of a process to generate FHIR XML bundles from HL7 messages, I have a subtransform transforming segments of an MDM_T02 message into a section of XML (EnsLib.EDI.XML.Document). So far this has worked well although I've encountered strange behaviour when making use of an xmlns attribute, whereby the attribute's name is duplicated despite the schema and DTL editor displaying it correctly.

Here's a snippet of the div.xmlns definition from the XSD schema used

Has anyone else encountered this? Any help is much appreciated

I'm attempting to import all the FHIR XSD files provided by http://hl7.org/fhir but get the immediate and unhelpful error message:

I've tried different versions of EOLs to no avail. Any idea what this error could mean?

Here's the XSD I'm trying to import:

I'm attempting to build a fairly complicated object graph, with nested objects and collections, in order to create a FHIR bundle.

In its most basic form there is a <bundle></bundle> element that represents the root, and there can be a number of nested <entry></entry> child elements.

I'm attempting to represent this as such:

class Bundle (%RegisteredObject, %XML.Adaptor)
{
    Property Entries As List Of Entry;
}

I'm trying to project a series of objects and nested objects to XML, and I'm having difficulties getting the XML representation exactly the way I need it to be. Here's what I'm working with, nothing complicated:

Class XMLTest Extends (%RegisteredObject, %XML.Adaptor)
{
    Property Id As Id;
}

Class Id Extends (%RegisteredObject, %XML.Adaptor)
{
    Property Value As%String (XMLPROJECTION = "attribute");
}

By default it seem closing tags are used with all elements, as such:

<XMLTest><IdValue="D1949EE1-7D03-4666-9548-D1A949E10327"></Id></XMLTest>

I have a business process that adds data to a global variable on receipt of an HL7 message, and a scheduled task that executes a class method defined within the same business process that removes data from the same global variable. With this in mind it makes sense to consider concurrency and therefore make use of the LOCK command.

My first question is whether this is actually necessary?

Hi

Is there a way to generate an MD5 checksum of a file by performing a checksum against the file's parts? We potentially have to process large files and therefore cannot keep the entire file in memory. From what I understand checksum related logic is typically stateful, allowing file data to be passed through sequentially, however it appears cache script offers only stateless functions offering a simple method of generating a single checksum hash from a single string (or stream), correct me if I'm wrong?

Hi

We have a use case whereby very large HL7 MDM messages, which contain many OBX segments, need to be processed by a custom business process. Part of the process is to iterate over each OBX segment and set a value in the same field for each segment.

Testing with a message that is ~500MB and contains over 8000 OBX segments I've encountered <STORE> errors, which surprisingly seems to be due to using SetValueAt. It appears to get through a few thousand segments before it falls over with the memory error, which I assume is because the SetValueAt method keeps in memory each preceding OBX segment?