< character in a 2.x OBX.5 (value) field

[fubar]

I've searched the faq and the entire site but was unable to find any hint about the following situation:

I'm trying to parse ORU R01 HL7 2.x messages with mirth 1.7

These are laboratory results in free text, and of course some OBX have OBX.5 (value) string elements like
" * CEFTAZIDIME <=8 S <=8 S"
It seems (after much fiddling!) that the mirth inbound message template parser will reject any message containing the "<" character no matter what I put for OBX.2 (value type) - I've tried TX and ST that I thought might allow those characters.

The ">" character is permitted, so I'm guessing it's not some confusion about XML delimiters.

I may be able to convince the people generating the message to do something about this, but I'm wondering if there's some value type I didn't yet try that will convince the parser to ignore any < characters inside an OBX.5?

Post edited by: fubar, at: 03/15/2008 13:32

Post edited by: fubar, at: 03/15/2008 13:33

Post edited by: fubar, at: 03/15/2008 13:33

Post edited by: fubar, at: 03/16/2008 10:47

[jacobb]

Re:&lt character in a 2.x OBX.5 (value) field

Turn on Encode Entities. The option is in your source and destination transformers, under the inbound and outbound template protocol properties.

[fubar]

Re:aracter in a 2.x OBX.5 (value) field

jacobb wrote:

Turn on Encode Entities. The option is in your source and destination transformers, under the inbound and outbound template protocol properties.

Thank you very much Jacob - absolutely beautiful solution. Works fine now.

Sorry I didn't notice that check box - I found the line end conversion but didn't stop to think that encode might do the trick. We always get lots of stupid characters like < in lab results and I would be surprised if it wasn't a common "gotcha"?

I wonder if it should always be on by default for 2.x messages - or is this usually not needed?

Post edited by: fubar, at: 03/15/2008 22:15

[afterdark23]

Re:&lt character in a 2.x OBX.5 (value) field

jacobb:
What does the Encode Entities do?


Does the Encode Entities need to be applied to every channel for this to work.

I have a LLP listener that feeds numerous channels and I am having this same problem. Can I select the Encode Entities in the main channel and that would take care of the other channels?


Also, there is an issue open in the Jira about this problem:
http://www.mirthproject.org/communit...owse/MIRTH-804

Post edited by: afterdark23, at: 03/17/2008 07:36

[jacobb]

Re:&lt character in a 2.x OBX.5 (value) field

Encode entities is on by default EXCEPT in version 1.7.0. This has been changed for 1.7.1, where there isn't even an option to turn it off.