HL7 Field Mapping…

This topic has 16 replies, 4 voices, and was last updated 15 years, 10 months ago by Tom Rioux.

Creator

Topic
March 16, 2009 at 7:14 pm #50721
Tom Rioux
Participant
Has anyone attempted to write a script or proc that would take the HL7 field mapping and change each field within an xlate according to the following example:

0(0).PID.00106.[0] to 0(0).PID.#3.[0]

I like using the # better than having the 5 digit numbering scheme. I would like to find an easy way to change the xlate without having to go in and touch every field. I understand that there may me user-defined segments (Z segments) that would not be able to get updated necessarily, but if it could get the bulk of it, that would be great.

Tom
Creator

Topic

Viewing 15 reply threads

Author

Replies
- March 16, 2009 at 9:46 pm #67235
  Jim Kosloskey
  Participant
  Tom,
  
  Let us know when you have that written and tested 😆
  
  I suspect that will be a non-trivial task as the position of the field is based on the position in a segment.
  
  So one would need to look at the particular variant and determine where within the message the segment is and where within the segment the field is.
  
  Since the same field can reside in different segments and the same segment can actually reside within different groups within the same message that could be a challenge. Especially challenging might be determining how to navigate the Cloverleaf(R) variant definitions.
  
  I have some ideas – but I think it is just easier to modify the Xlate as I need to. In 5.6 the relative number notation is the default for new Xlates.
  
  email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.
- March 17, 2009 at 2:01 pm #67236
  Tom Rioux
  Participant
  Hey Jim,
  
  I started working on something yesterday afternoon for a bit. I was attempting to use the the “fldmap” file from the hl7 directory under the root. However, after studying the file and all of the xlates and hl7 variants, I realized that there are too many variables to consider when attempting to do this. I thought of everything you mentioned and then some. I don’t think my tcl/scripting skills are accomplished enough to do this justice. It would take a power far, far greater than me (Charlie! 😀 ) to get something of this magnitude done.
  
  Tom
- March 19, 2009 at 2:18 pm #67237
  Charlie Bursell
  Participant
  Looked like a bit of a challenge so, since I had a couple of hours free yesterday I came up with this.
  
  USAGE:
- March 19, 2009 at 5:39 pm #67238
  Tom Rioux
  Participant
  Charlie,
  
  This proc is awesome. Thanks so much. I did make one adjustment to it after downloading it. I kept getting the following error:
  
  Serious Error!! Offset not found in Segments array – ABORTED!!!”
  
  It appears as though that it had something to do with a possibly the field notation (i.e. in this case 90058) not having an offset in one of the variants that is being used. So, what I did was commented that part out and put a continue in there. In those cases where it would have normally aborted, it left the original field notation in place. These are few and far between and can be adjusted manually as needed. Does what I did sound right according to the logic in your script? I tried putting echo’s out to make sure I knew what I was seeing come in and go out.
  
  Thanks again…Tom
- March 19, 2009 at 7:11 pm #67239
  Charlie Bursell
  Participant
  What you did will keep the script from aborting but what about the undefined field? That was the reason I aborted at that point to let you know you had a malformed variant.
  
  I pull the name of the segment and the field address from the statement.
  
  I then look for that segment first in the defined variant. If the segment is not found there I pull it from the root. I then number the fields from 1 – x
  
  If that field is not defined in the segment using it, it is a problem that should be fixed.
  
  I guess I could issue a warning and simply skip that file.
  
  With that said I would not be surprised if you find troubles I tested it against about 25 – 30 Xlates I had.
- March 19, 2009 at 7:33 pm #67240
  Tom Rioux
  Participant
  Charlie,
  
  This particular Xlate I’m testing with has a different inbound and outbound variants. The segment size is larger in the inbound than in the outbound. It is a Z-segment and is pretty much only used to read from inbound. We don’t do hardly any mapping to this segment outbound. I’m assuming that since the field exists in the inbound variant and the not in the outbound variant, that is why it is throwing the error.
  
  Would this be considered a malformed variant? Should I go in and update all of the variants that have this type of issue?
  
  Thanks for all your help!!!
  
  Tom
- March 19, 2009 at 7:50 pm #67241
  Charlie Bursell
  Participant
  No, the script uses the OB variant for OUT and tilde prefixes. It use the IB variant for all else
- March 19, 2009 at 7:54 pm #67242
  Charlie Bursell
  Participant
  Tom:
  
  If you send me the Xlate file and variant in a zip I will test it. Name the zip like zpp or something insteado zip as our lovely IT guys do not allow zips
  
  cbursell@healthvision.com
- March 19, 2009 at 11:47 pm #67243
  Charlie Bursell
  Participant
  Tom Rioux pointed out a problem.
- March 24, 2009 at 6:44 pm #67244
  Tom Rioux
  Participant
  Charlie,
  
  It appears I may have run into another issue. I’m attempting to go through your code but my head hurts trying to understand it all. I bow to the master.
  
  ANYWAY….what I’m seeing, when I run the script against some xlates, is that it doesn’t appear to reaching into nested IF statements or IF statements within iterates.
  
  Can you check that xlate I sent you? If you run the script against and then check a segment like the AL1 segment. When I ran it, it didn’t change any of the notations in that segment. Hopefully, I’m not using an older copy. I will download the script again and try it again.
  
  Thanks…Tom
- March 24, 2009 at 7:11 pm #67245
  Charlie Bursell
  Participant
  When I get one of those round tuits, I’ll look at it
- March 24, 2009 at 7:59 pm #67246
  Tom Rioux
  Participant
  I know what you mean. In between working on xlates and getting ready for go-lives, boss has laid some round tuits on my desk that are labeled “documentation”. I hate those kind of round tuits but they are a necessary evil. 😈 😥
  
  Tom
- March 24, 2009 at 8:22 pm #67247
  Max Drown (Infor)
  Keymaster
  dokuwiki has made my documentation life soooooooo much easier.
  
  -- Max Drown (Infor)
- March 24, 2009 at 10:03 pm #67248
  Charlie Bursell
  Participant
  I guess we will keep trying until we get it right 😀 Please remember that I am doing all of this pro bono 0 no pay 😥
  
  The problem was not in parsing it was in the regular expression that extracted the field addresses (addrPat). It did not take into account iterations like (%s1), etc. It was looking for all numbers. Besides the regexp was much more complicated than it should be.
  
  Try the new one
- March 24, 2009 at 10:13 pm #67249
  Tom Rioux
  Participant
  I understand about the pro bono and do thank you very much for all of the work you have done on this. I’m sure many others are gonna take advantage of this as well. I will test this out and let you know.
  
  Tell you what Charlie…I’ll buy your next round of golf!! 😀
- March 24, 2009 at 10:40 pm #67250
  Tom Rioux
  Participant
  THANKS CHARLIE! WORKS LIKE A CHAMP!!!
Author

Replies

Viewing 15 reply threads

The forum ‘Cloverleaf’ is closed to new topics and replies.