How can I pass a parameter to a symbolic field?

Wilhelm

Try using a global labelled common area or @USER variable.

Either use persistent @USER variables initialized in a stored procedure or use Labelled COMMON setup the same way. It then depends on where the Symbolic is to be used. In Reports (R/LIST, etc.), you initialize the variables before the routine to do the report is called and clear them after the report ends. If you're using them from within a Form, then initialize them in the READ event and clearing them on the SAVE or CLEAR event if their scope is limited to a single record. Otherwise, initialize the data on the CREATE event and clear them on exit.

Don M.

Don (and Mattew)

sorry… I forgot to mention, that we are working with remote (dynamic) engines. The R/List request from the (JOI) client comes in through OECGI, passes inet_xml and after this, a remote engine on the server is activated for the required processing.

Since we don't have permanent sessions, the common labeled area is restricted to essential information, that is included only in the particular request.

The problem behind the question is this:

We have user defined input forms, based on associated multivalues. This way, there is only 1 mv-field (something like "LABEL") in the dictionary and 1 mv-field for the actual data (something like "INPUT_VALUE". (For validation, popups, oconv etc. we have additional mv-fields.) Each form may have a completely different data structure.

As an example: In the user-defined-input-form 1 (UDIF1), the label "TRANSACTION _DATE" could be in row position 5, while the same label in UDIF2 might be in row position 8. In UDIF3 there might be no such label at all, etc…

To filter certain documents for the existence of label "TRANSACTION_DATE" I would like to issue an R/LIST command similar to "LIST DOCUMENTS LABEL INPUT_VALUE WITH TRANSACTION_DATE=THIS_WEEK() BY … ".

I could write a symbolic, for this specific occurrence. But imagine the nightmare with hundreds or thousands of UDFs - and worst of all, each new label would require dictionary maintenance, which wouldn't be user friendly at all.

We thought of adding an MFS, but isn't there another way?

Any suggestions are welcome.

- Wilhelm -

Do the term UGLY mean anything? . I'm glad it's you and not me. I would have shoved the data out to a dataset or file and let some other process handle it .. or I might have said "oops" and taken a vacation.

Don M.

Wilhelm,

Can a symbolic that prefixes values with labels be of any help?

Say LABEL has the value(s) "NAME" :@VM: "AGE", and INPUT_VALUE the value(s) "Bob" :@VM: "33". Then the symbolic LABELLED_VALUE I'm thinking of would be something like "NAME:Bob" :@VM: "AGE:33".

- Oystein -

Oystein,

Thanks for the idea. I'll give it a try.

Regards

Wilhelm

Don,

…I would have shoved the data out to a dataset or file….

If the user must be able to create his own templates, you can't know the data structure in advance.

On the other hand, there is practically no processing done with the input - except storage and retrieval. The important point is to filter the occurrence of certain values and labels and list them.

- Wilhelm -

Wilhelm,

Additional thoughts:

(1) You might have labels with values that need conversion or justification for sort and queries to work properly. Here's an example of a solution for integer values:

LABEL=AMOUNT"

INPUT_VALUE=450"

LABELLED_VALUE=AMOUNT:000450".

Of course now your symbolic's formula needs to know more about each label - which conversions to apply for which labels. I can think of several ways to solve that problem:

- If conversion is only relevant for a few labels let the symbolic have a list of those labels, along a list of which conversions to apply

- Let the labels have names that imply which conversion to apply, e.g, I6_AMOUNT for 6 digit integers

- Etc.

(2) You can have XREF like values too:

LABEL=TEXT"

INPUT_VALUE=Big bada boom"

LABELLED_VALUE=TEXT:Big" : @SVM : "TEXT:bada" : @SVM : "TEXT:boom"

- Oystein -

View this thread on the forum...