Re: [HAPI-devel] How are segment names and their index generated?

[Jonathan B. <jon...@mi...>]

It does help. Now I understand how/why I’m seeing what I’m seeing so I can
work with it.



For what it’s worth, I’ve been asking questions about HAPI because I’m
working on moving a complex channel from Mirth Connect (which uses HAPI) to
its own Java code. There is a different use case and some extra plumbing in
Mirth Connect that almost-but-not-quite matches with HAPI and its throwing
me off a bit.



 Thanks James.



*From:* jam...@gm... [mailto:jam...@gm...] *On Behalf Of *James
Agnew
*Sent:* Friday, October 22, 2010 3:02 PM
*To:* Jonathan Bartels
*Cc:* hl7...@li...
*Subject:* Re: [HAPI-devel] How are segment names and their index generated?



Hi Jonathan,

A bit of background:

When a particular structure has more than one of the same segment in a row,
according to HL7, those are repetitions. An obvious example would be the
NTEs attached to an OBX.

On the other hand, when a structure has a segment of the same type but in
different locations within the structure, these are treated as distinct
segments. An obvious example of this would be the A24 "link patients"
structure, which exactly two PID segments, optionally separated by a few
related segments (PD1, PV1, etc).

Because in this case these segments have the same name (in this case PID)
but a different interpretation (in this case "Patient 1 Identification" and
"Patient 2 Identification") HAPI needs a unique name for each, so it
accomplishes this by appending a number to the end.

Now, when it comes to a GenericMessage, HAPI doesn't actually know the
"correct" structure for the message it is parsing, so it has to try and
infer it from the actual segments it finds. The logic it uses is to treat
multiple segments in a row as repetitions, but multiple segments separated
by other segments as unique segments. This behaviour also allows HAPI to be
able to encode the message back in the same order it received it.

My advice to you would probably be to call
String[] names = message.getNames();
And then iterate over that list. If you have a name start starts with "NTE",
you know it's an NTE, and you can infer which segment it belongs to using
the previous name.

Hope this helps!

James

On Fri, Oct 22, 2010 at 1:18 PM, Jonathan Bartels <jon...@mi...>
wrote:

I’m having a problem using a Terser on a GenericMessage with repeating NTE
 segments.



I whipped up a test case to show the issue. Why does the terser getter like
“NTE2-3-1” work while “NTE(1)-3-1” fail? Is there any way to make the latter
format work?



@Test

public void testNTEQuirk() {

                String hl7Message = null;

                try {

                                hl7Message =
resourceUtil.getResourceAsString("nte_quirk.hl7");

                } catch (IOException e) {

                                fail("Could not load resource!");

                }



                GenericMessage message = new
GenericMessage.UnknownVersion(new DefaultModelClassFactory());



                try {

                                message.parse(hl7Message);

                } catch (EncodingNotSupportedException e) {

                                e.printStackTrace();

                                throw new RuntimeException("Could not parse
HL7 message! Encoding not supported.", e);

                } catch (HL7Exception e) {

                                e.printStackTrace();

                                throw new RuntimeException("Could not parse
HL7 message!", e);

                }

                Terser terser = new Terser(message);



                //"/." + segment + "(" + index + ")-" + field + "-" +
component;

                try {

                                //why do the next 3 lines work??

                                assertEquals("OBRnotes",
terser.get("/.NTE2-3-1"));

                                assertEquals("OBX",
terser.get("/.NTE3-3-1"));

                                assertEquals("notes",
terser.get("/.NTE4-3-1"));

                                assertEquals("ORCnotes",
terser.get("/.NTE(0)-3-1"));

                                //why do these lines fail??

                                assertEquals("OBRnotes",
terser.get("/.NTE(1)-3-1"));

                                assertEquals("OBX",
terser.get("/.NTE(2)-3-1"));

                                assertEquals("notes",
terser.get("/.NTE(3)-3-1"));

                } catch (HL7Exception e) {

                                fail();

                }

}





MSH|^~\&|ADM|CLF654321||CLF123456|200612131130||ORU^R01|4323424|P|2.5|||AL|NE

PID|||||

ORC|||||||||||||||||||

NTE|1||ORCnotes|

OBR||||||||||||||

NTE|2||OBRnotes|

OBX|1|||||||||||||||

NTE|3||OBX|

NTE|4||notes|


------------------------------------------------------------------------------
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev
_______________________________________________
Hl7api-devel mailing list
Hl7...@li...
https://lists.sourceforge.net/lists/listinfo/hl7api-devel