CL 6.0: subcomponent reference issue

Clovertech Forums Read Only Archives Cloverleaf Cloverleaf CL 6.0: subcomponent reference issue

  • Creator
    Topic
  • #53809
    Jim Kosloskey
    Participant

      Coming from Clovelreaf 5.6 to 6.0.

      Two issues to report.

      First –

      In CL 5.6 when a field contained an erroneous subcomponent separator (typically the dreaded &) a reference to the component without subcomponent notation would result in the first piece of data being parsed. For example say the first component of a field has the vale ‘A&B’. When referencing the first component only ‘A’ would be parsed.

      In CL 6.0 the entire data is now available so suddenly where once one got ‘A’ as a result of a component notated COPY in an Xlate now one gets ‘A&B’. That is good that now the entire data is available BUT it does not give the same results for existing integrations unless adaptations are made.

      Second –

      In Cl 6.0 if a reference is made to a field at a higher level (say at the field or component level) and then later in the Xlate another reference is needed at the subcomponent level, the subcomponent notation does not work.

      Example:

      a COPY of …NK1.#2(0) –> @junk followed by

      a COPY of …NK1.#2().[1].[0] –> some destination

      Assume the value in the second component of NK1-2 being ‘A&B’.

      This hass 2 subcomponents and the second COPY above is trying to get the first subcomponent only but waht will be in ‘some destination’ is ‘A&B’.

      To get what is needed – this works:

      a COPY of …NK1.#2(0).[0].[0] –> @junk (note the component and subcomponent notation) followed by

      a COPY of …NK1.#2().[1].[0] –> some destination

      That will result in ‘some destination’ getting the value ‘A’.

      This has been reported to Infor.

      email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

    Viewing 10 reply threads
    • Author
      Replies
      • #79013
        Lynda Chen
        Participant

          Has Infor responded to this discovery with a fix by the CL 6.1 release?

        • #79014
          Jim Kosloskey
          Participant

            Infor has not responded to our report.

            However, I have discovered in later than 6.0.0 releases of Cloverleaf the parsing has returned to what we experienced in 5.6. So perhaps in a way Inoor has responded.

            So special handling needs to occur if one wishes to change the offending subcomponent seperator (maybe to th standard T or ???).

            I have a proc which does that.

            email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

          • #79015
            Bob Richardson
            Participant

              Greetings,

              According to the Release Notes for Integrator 6.0.1 the sub-comonent problem is fixed.  We are planning to upgrade from 5.8.7 to 6.1.0 hoping to get the .01 patch beforehand.

              I have found that one must read ALL the Release Notes from the base release version (in this case 6.0) thru whatever level you plan to upgrade to in order to find all the details on fixes, enhancements, etc.

              Enjoy.

            • #79016
              Lynda Chen
              Participant

                We are working on upgrading from CL 5.8.7 to CL 6.1. Found that CL 6.1 is not removing trailing subcomponent separators even when empty.

                Thus in CL 5.8

                incoming: |Lastname^Firstname^MI^^^MD&&|

                outgoing from xlt with a copy: |Lastname^Firstname^MI^^^MD|

                in CL 6.1

                incoming: |Lastname^Firstname^MI^^^MD&&|

                outgoing from xlt with a copy: |Lastname^Firstname^MI^^^MD&&|

              • #79017
                Bob Richardson
                Participant

                  Greetings,

                  Well,  looks like they may have introduced a new problem!

                  Please submit an incident at Infor’s website.

                  Maybe they can fix this in the planned .01 patch!

                  Thanks for reporting this issue to us.

                  Enjoy.

                • #79018
                  Jim Kosloskey
                  Participant

                    Hmmm it has been a while since I studied thiis but I think in HL/7 parsing (not Cloverleaf specific) trailing subcomponent characters are supposed to exist.

                    So maybe this release actually more tightly follows the standard.

                    Again it has been some time since I studied that specific rule so I coulld be way off base here.

                    email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

                  • #79019
                    Bob Richardson
                    Participant

                      Greetings,

                      Jim, you are correct from what we see here regarding HL7 standards and enforcing the rules.  The 2.6 variants now include deprecated fields that are INSERTED (!) in several segments throwing off our field alignments.

                      We fixed them by creating the segments under 5.8.7 (like the IN1) and then migrating to 6.1 to avoid dragging in the new definitions when the variant segments do not exist in our 2.6 variants.

                      However, like empty trailing components, I would expect the XPM parser to not include trailing empty sub-components too.

                      We will be running 6.1 in our Test environment next week and if not posted as an incident we certainly could do that based on test results.

                      Thanks for chiming in here.

                    • #79020
                      Lynda Chen
                      Participant

                        FYI: Infor has accepted the issue of the trailing sep chars and will try to work on removing them.

                        Another difference we noticed was:

                        CL 5.8 engine passes components and subcomponents as a list to tclprocs.

                        However, with CL 6.1 the engine continues to pass components as lists, but subcomponents as a string to tclprocs.

                      • #79021
                        Jim Kosloskey
                        Participant

                          Lynda,

                          Coud you email me – I would like to discuss this finding with you?

                          email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

                        • #79022
                          Bob Richardson
                          Participant

                            Greetings,

                            Would it be possible to post an example of the Xlate or tclproc code

                            that you are using in 6.1 and 5.8?  that is, of the sub-component problem?

                            We are going to be migrating from 5.8.7 to 6.1 very soon now and are compiling a list of items to test/evaluate.

                            Thank you.

                          • #79023
                            Lynda Chen
                            Participant

                              Here are Notes added by R&D on 17 Dec 2014 09:27:41 :

                              We plan to fix the trailing sep char handling in CIS6.1.1.0, after the fix, engine will remove these trailing sep chars as what it did before.

                              The current behavior that cloverleaf handles component and subcomponent in tcl is the result of the HL7 subcomponent copy change. As mentioned, after the change, if user uses a component level address in COPY, cloverleaf will get the whole component content (not expand it into list), which is used for COPY source to solve the 5.8 copy issue.

                              There are 2 workarounds here:

                              1. Use sub-component level address as source in COPY

                              2. Split the xlateInVals by sub-component sep char in tcl pre/post proc.

                              Unfortunately, the new behavior is the base of the fix.

                          Viewing 10 reply threads
                          • The forum ‘Cloverleaf’ is closed to new topics and replies.