Reusable TRANS I-descriptor dictionary items.

Have you ever created an I-descriptor for a "one-off" report only to
find that you've now created a demand for this "new" piece of
information?  With a little extra work, these I-descriptors can be made
flexible enough to work in more than one dictionary.  A sample approach
might go like this:

Suppose we are working on a checkout statistics problem.  We may need to
relate three types of data: demographic, bibliographic, and transaction
information.  These data are related to PATRON, B.MASTER, and CHECKOUT
files, and are keyed by the patron barcode, LCN, and pieces barcode
respectively.  If we create an I-descriptor, say L.B.MASTER.KEY, in each
file that we run reports from, we can use any I-descriptor that performs
a TRANS based on B.MASTER.KEY.

We can create keys involving complex inter-relationships between files
by building them from simpler keys.  Begin by building keys for the
three types of data, in each file that may be used to originate a
report.  Note that each set of I-descriptors will be identical for any
file with the same key (@ID).

To review, here is a typical I-descriptor:
DICT PATRON L.PIECES.KEY                 - L. prefix denotes local item
0001: I                                  - Indicates I-descriptor
0002: TRANS("BLOCKS",PATRON.KEY,9,"X")   - Field definition
0003:                                                - Conversion code (optional)
0004: Item bar code                      - Column heading (optional)
0005: 14L                                - Output format
0006: S                                  - Single/Multivalued indicator

To save room only the item name and attribute 2 are shown below:

File:   PATRON (or BLOCKS).
L.PATRON.KEY       @ID
L.PIECES.KEY       TRANS("BLOCKS",L.PATRON.KEY,9,"X")
L.B.MASTER.KEY     TRANS("PIECES",L.PIECES.KEY,1,"X")

File:   CHECKOUT (or PIECES):
L.PIECES.KEY       @ID
L.B.MASTER.KEY     TRANS("PIECES",L.PIECES.KEY,1,"X")
L.PATRON.KEY       TRANS("CHECKOUT",L.PIECES.KEY,5,"X")

File:   B.MASTER  (or B.MARC)
L.B.MASTER.KEY     @ID
L.PIECES.KEY       TRANS("B.MASTER",L.B.MASTER.KEY,3,"X")
L.PATRON.KEY       TRANS("CHECKOUT",L.PIECES.KEY,5,"X")

Now we can create some "universal" I-descriptors containing information
in any of these files, that will be identical in each file, for example:

L.PATRON.CODE      TRANS("PATRON",L.PATRON.KEY,14,"X")
L.COLL.CODE        TRANS("PIECES",L.PIECES.KEY,8,"X")
L.CHKOUT.SUBLOC    TRANS("CHECKOUT",L.PATRON.KEY,14,"X")
L.PATRON.NAME      TRANS("PATRON",L.PATRON.KEY,1,"X")
L.TITLE
   OCONVS(TRANS("B.MASTER",L.B.MASTER.KEY,4,"X"),"TA.MASTER;X1;;1")
L.AUTHOR
   OCONVS(TRANS("B.MASTER",L.B.MASTER.KEY,5,"X"),"TA.MASTER;X1;;1")
L.SUBJECT
   OCONVS(TRANS("B.MASTER",L.B.MASTER.KEY,6,"X"),"TA.MASTER;X1;;1")
Note that the last three use OCONV to extract only the first subvalue of
the authority.

Now, to list titles checked out to each patron with checkout
sub-location:
    SELECT CHECKOUT SAVING UNIQUE L.PATRON.KEY
    SORT PATRON BY L.PATRON.NAME L.PATRON.NAME L.TITLE L.CHKOUT.SUBLOC
or alternatively:
    SORT CHECKOUT BY L.PATRON.NAME L.PATRON.NAME L.TITLE L.CHKOUT.SUBLOC
or:
    SELECT CHECKOUT
    SORT PIECES BY L.PATRON.NAME L.PATRON.NAME L.TITLE L.CHKOUT.SUBLOC

A useful way to output these translations is in a "spreadsheet" format.
To count checkouts by sub-location and patron group, create additional
I-descriptors for each collection breakdown you're interested in:

L.COLL.RBR   IF (L.COLL.CODE EQ "RBR") THEN "1" ELSE "0"
L.COLL.A.V   IF (L.COLL.CODE EQ "A/V") THEN "1" ELSE "0"
L.COLL.OTHER
   IF (L.COLL.CODE NE "RBR") AND (L.COLL.CODE NE "A/V") THEN "1" ELSE
"0"

Now, to get a "spreadsheet" of collection vs. sub-location:
SORT CHECKOUT BY L.CHKOUT.SUBLOC BREAK-ON L.CHKOUT.SUBLOC TOTAL
     L.COLL.RBR TOTAL L.COLL.A.V TOTAL L.COLL.OTHER ID-SUPP DET-SUPP
     HEADING "Circ stats spreadsheet'L'"