dictionary not working

Assuming I understand your F-correlative correctly (but it is a long time since I used either A or F correlatives), I would use something like this:

First, create a field defining your attribute 19:

DICT BS.TEMP PL.IDS
1: D
2: 19
3:
4: PL ids
5: 7R
6: M
7: A-PL

Now, create an I-type that splices the order id onto those ids:

DICT BS.TEMP ORDER.PL.IDS
1: I
2: SPLICE(REUSE(@ID), '*', PL.IDS)
3:
4: Order PL ids
5: 12L
6: M
7: A-PL

Now, list your file:

LIST BS.TEMP PL.IDS ORDER.PL.IDS
BS.TEMP...   PL ids.   Order PL ids
1                001   1*001
                 002   1*002
                 002   1*002

Note: Both dictionaries are defined a "multi-valued" (field 6) and have an "association" defined (field 7) so the query processor knows that the multivalues in these fields are related.

I hope this helps.

Brian

DICT BS.TEMP PL.IDS
1: D
2: 19
3:
4: PL ids
5: 7R
6: M
7: A-PL

DICT BS.TEMP ORDER.PL.IDS
1: I
2: SPLICE(REUSE(@ID), '*', PL.IDS)
3:
4: Order PL ids
5: 12L
6: M
7: A-PL

LIST BS.TEMP PL.IDS ORDER.PL.IDS
BS.TEMP...   PL ids.   Order PL ids
1                001   1*001
                 002   1*002
                 002   1*002

This works perfectly.  As we came from the D3 world, we were used to A and S items. We have only used a few I dictionary items on anything new.

I made two new dictionary items from your example.

and then

now I select my WO file

and list my PL file with these new keys

perfect.

Thanks

Hi Jesse,

My advice would be to embrace I-types for your calculations. This is partly because QM does not fully support A/S types (e.g. LPV is not supported), but mostly because the expressions in I-types are much more readable than A or F correlatives.

We started in MV using the ADDS products, then mvBASE, then UniVerse, and now QM (over a 40 year period). Along the way, we have migrated most of our dictionaries to D and I-types, and many queries are much more readable and sensible than they were in the PICK days.

If you define your dictionaries nicely, you can use the GENERATE command to generate an INCLUDE item for your QMBasic programs that define the record the layout for each file. Then you can use those included tokens in your programs, and the programs (a) become more readable; and (b) you have less chance of specifying an incorrect attribute number when accessing a record.

The ITYPE function is also great for getting a value from the database from within your programs.

Feel free to contact me if you have any (more) questions.

Cheers,

Brian