Newbie question regarding using fields from a table that is brought in twice in a synonym created through the meta data canvas.

I am trying to bring in email addresses from a table goremal that contains lots of emal codes. I am bringing the table in twice to the synonym and specifying which email code to use for each in the join properties of the synonym. When I bring the table in once the report works fine. When I bring the table in a second time the report gives me the error in the subject field.

Below are the output from the initial procedure when I get the correct email out. Then I posted the error message. Below that is the coding for the second procedure when I am have added in the second goremal table and tried reporting out information from it.

TABLE FILE STUDENT
BY HIGHEST STUDENT.SPRIDEN.SPRIDEN_ID
BY HIGHEST STUDENT.SPRIDEN.FULLNAME_BANNER_FUNCTION
BY STUDENT.SPRIDEN.SPRIDEN_FIRST_NAME
BY STUDENT.SPRIDEN.SPRIDEN_LAST_NAME
BY HIGHEST 1 STUDENT.GOREMAL_1.GOREMAL_ACTIVITY_DATE
BY STUDENT.GOREMAL_1.StuEmail
BY STUDENT.SGBSTDN.SGBSTDN_STST_CODE
BY STUDENT.SGBSTDN.SGBSTDN_LEVL_CODE
BY STUDENT.SGBSTDN.SGBSTDN_STYP_CODE
WHERE ( STUDENT.SYVPER1.SYVPER1_SEX EQ &SYVPER1_SEX.(OR(FIND STUDENT.SYVPER1.SYVPER1_SEX IN STUDENT)).Gender. ) AND ( STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_1 EQ &SGBSTDN_MAJR_CODE_1.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_1 IN STUDENT)).Major. ) AND ( STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_CONC_1 EQ &SGBSTDN_MAJR_CODE_CONC_1.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_CONC_1 IN STUDENT)).Concentration. ) AND ( STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_MINR_1 EQ &SGBSTDN_MAJR_CODE_MINR_1.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_MINR_1 IN STUDENT)).Minor. ) AND ( STUDENT.SGBSTDN.SGBSTDN_STST_CODE EQ &SGBSTDN_STST_CODE.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_STST_CODE IN STUDENT)).Status. ) AND ( STUDENT.SGBSTDN.SGBSTDN_LEVL_CODE EQ &SGBSTDN_LEVL_CODE.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_LEVL_CODE IN STUDENT)).Level. );
WHERE STUDENT.SPRIDEN.ChangeInd;
ON TABLE SET PAGE-NUM NOLEAD
ON TABLE SET ASNAMES ON
ON TABLE NOTOTAL
ON TABLE PCHOLD FORMAT HTML
ON TABLE SET HTMLCSS ON
ON TABLE SET STYLE *
INCLUDE = IBFS:/EDA/EDASERVE/_EDAHOME/ETC/endeflt.sty,
$
ENDSTYLE
END
--------------------------------------------

0 ERROR AT OR NEAR LINE 25 IN PROCEDURE ADHOCRQ FOCEXEC * (FOC029) ALL SORT KEYS ARE NOT IN A SINGLE TOP-TO-BOTTOM SEGMENT PATH
---------------------------------------------



TABLE FILE STUDENT
BY HIGHEST STUDENT.SPRIDEN.SPRIDEN_ID
BY HIGHEST STUDENT.SPRIDEN.FULLNAME_BANNER_FUNCTION
BY STUDENT.SPRIDEN.SPRIDEN_FIRST_NAME
BY STUDENT.SPRIDEN.SPRIDEN_LAST_NAME
BY HIGHEST 1 STUDENT.GOREMAL_1.GOREMAL_ACTIVITY_DATE
BY STUDENT.GOREMAL_1.StuEmail
BY HIGHEST 1 STUDENT.GOREMAL.GOREMAL_ACTIVITY_DATE
BY STUDENT.GOREMAL.HomeEmail
BY STUDENT.SGBSTDN.SGBSTDN_STST_CODE
BY STUDENT.SGBSTDN.SGBSTDN_LEVL_CODE
BY STUDENT.SGBSTDN.SGBSTDN_STYP_CODE
WHERE ( STUDENT.SYVPER1.SYVPER1_SEX EQ &SYVPER1_SEX.(OR(FIND STUDENT.SYVPER1.SYVPER1_SEX IN STUDENT)).Gender. ) AND ( STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_1 EQ &SGBSTDN_MAJR_CODE_1.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_1 IN STUDENT)).Major. ) AND ( STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_CONC_1 EQ &SGBSTDN_MAJR_CODE_CONC_1.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_CONC_1 IN STUDENT)).Concentration. ) AND ( STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_MINR_1 EQ &SGBSTDN_MAJR_CODE_MINR_1.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_MAJR_CODE_MINR_1 IN STUDENT)).Minor. ) AND ( STUDENT.SGBSTDN.SGBSTDN_STST_CODE EQ &SGBSTDN_STST_CODE.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_STST_CODE IN STUDENT)).Status. ) AND ( STUDENT.SGBSTDN.SGBSTDN_LEVL_CODE EQ &SGBSTDN_LEVL_CODE.(OR(FIND STUDENT.SGBSTDN.SGBSTDN_LEVL_CODE IN STUDENT)).Level. );
WHERE STUDENT.SPRIDEN.ChangeInd;
ON TABLE SET PAGE-NUM NOLEAD
ON TABLE SET ASNAMES ON
ON TABLE NOTOTAL
ON TABLE PCHOLD FORMAT HTML
ON TABLE SET HTMLCSS ON
ON TABLE SET STYLE *
INCLUDE = IBFS:/EDA/EDASERVE/_EDAHOME/ETC/endeflt.sty,
$
ENDSTYLE
END

This message has been edited. Last edited by: <Kathryn Henning>,

One of the fundamental rules of hierarchical structures (which you're creating with your joins in the metadata) is that the fields you sort on in your select need to follow a single path down through the hierarchy. You can't do a

sort on parent,
sort on left child,
sort on right child.

It doesn't actually make sense from a mathematical point of view.

Two ways to solve this --

1. Change your joins so that each segment pulls one child along behind it. That is, turn parent, left child, right child into parent, left child, left child's child. Have the joins put all three segments in a straight line instead of in a forking structure.

2. Take two passes at the data. Pass one builds a hold file with the Sort fields available in the parent and left child, then join the resulting hold file with the source again and select the Sort fields available in the right child.

Clear as mud? This is a pretty down-in-the-weeds concept.

Also do not forget that you can use CHECK FILE xxxxxx PICT after your JOIN to see a "representation" of your structure.

This will often help you understand the explanation that John has provided.

T

I guess then my followup question is...

since this is data that is to be used over and over again in quite a few applications and I need to get this flattened out structure of most recent email types and addresses all the time would it be better, performancewise, to create a new table in a database using WF with the flattened data in it. Then use this new table/synonym in my applications. I would update the table nightly through report caster. Any other suggestions on how best to do this would be greatly appreciated.

Sounds like you have the right plan worked out already!

If this is for a reporting application, you want to spend the time up front to make the maintenance of code as simple and inexpensive as possible. Flattening out the data structure into a dedicated table for the purpose is an excellent idea.

[CLOSED]