Problem with initializing an array with comp and packed data

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

Hi,

I want to initialize an array. This array contains comp and packed data and is part of a record structure..Explained below:

dick scherrer · Posted: Thu Sep 02, 2010 1:55 am

Hello,

Suggest you read here:
publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/handheld/Connected/BOOKS/igy3lr50/6.2.20.2

Ronald Burr · Posted: Thu Sep 02, 2010 2:09 am

Are you finding the "unexpected" values immediately after initialization, or only following one or more FETCHes?

If after a FETCH, it would help to see the CURSOR declaration along with the table structure (DCLGEN).

Bill O'Boyle · Posted: Thu Sep 02, 2010 2:09 am

Beginning with VS/COBOL II, if you add VALUE clauses to all array elementary items, they will be initialized during program prologue.

Otherwise, there's INITIALIZE (as Dick has pointed out) which is used during execution.

Bill

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

Refering to my first post...

WS-REC-7382 is the structure which I have declared in my pgm..

WS-T7382........is the DCLGEN structure which I have copied in my program

before fetch I am initializing both the structures::

INITIALIZE WS-REC-7382
WS-T7382.

The fetching data into the DCLGEN structure WS-T7382

Then finally MOVE WS-T7382 TO WS-REC-7382.

AND I am getting some unexpected values fter this move statement.

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

dick scherrer · Posted: Thu Sep 02, 2010 2:38 am

Hello,

dbzTHEdinosauer · Posted: Thu Sep 02, 2010 2:40 am

deleted by poster

Bill O'Boyle · Posted: Thu Sep 02, 2010 2:43 am

You can first move LOW-VALUES to WS-REC-7382 (will generate an MVCL Assembler instruction), then followed by -

Ronald Burr · Posted: Thu Sep 02, 2010 2:44 am

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

Well, WS-T7382 has the same structure as I had posted in my first post except the array part

01 WS-T7382.
05 TS-NBR PIC S9(4) USAGE IS COMP.
05 TS-RAN-NBR PIC S9(4) USAGE IS COMP.
05 WS-TS-CODE-ST-1 PIC S9(4) COMP.
05 WS-T6-PL-COUNT-1 PIC S9(9) COMP-3.
.......
.......
.......

05 WS-TS-CODE-ST-50 PIC S9(4) COMP.
05 WS-T6-PL-COUNT-50 PIC S9(9) COMP-3.
05 WS-ADD-UPD PIC X(08).
05 WS-CD-ENT PIC X(06).
.........
.........

My objective to define my own structure is to get the field WS-TS-CODE-ST-1 depending upon some subscript, so I have declared the array structure in pgm.

And regarding the data fetch it is no cursor but a simple select query fetching all columns of the table.

It is happening in a perform loop:
----------------------------------------

PERFORM UNTIL EOF

***
Initialize WS-T7382
WS-REC-7382

SELECT * into WS-T7382
from table
where..............

If success

MOVE WS-T7382 to WS-REC-7382
and do further processing.
endif

end-perform..

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

for each iteration of the perform the where cond will change and a single row from the table will be retrieved..more than one rows shall not be retrieved for sure so I have used no cursor..

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

dick scherrer · Posted: Thu Sep 02, 2010 3:26 am

Hello,

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

Yes, you are right to an extent..

I don't want to use dynamic sql that is the reason I am using this array structure in my program..I only need the field WS-TS-CODE-ST. And the subscript for this I will get dynamically whwn I run by reading another table.

So instead of putting 50 evaluates or dynamic sql I am trying the array approach..so that when I get the subscript value dynamically I can only retrieve the required information... WS-TS-CODE-ST (subscript)

In the DCLGEN I did not have a array structure so had to define it in my program and hence I am moving MOVE WS-T7382 to WS-REC-7382

Ronald Burr · Posted: Thu Sep 02, 2010 3:46 am

There is no way that performing a query 40 or 50 times with different predicate values is going to populate 40 or 50 DIFFERENT elements in an array ( Hint - there is no subscript/index on your host-variable name(s)) to show which array element(s) to put the results into.
1) Never a good idea to code SELECT * in a static SQL ( i.e. in a program ). One simple change to the table and you will blow up.
Try something like this (note: NOT tested):

dick scherrer · Posted: Thu Sep 02, 2010 3:50 am

Hello,

You have majorly confused yourself. . .

How did dynamic sql get into the picture? Nothing appears to be dynamic. . .

Why might you want 50 evaluates?

The only way to successfully move some data to an array item is to move the data to that particular item (using a subscript or index).

Suggest you forget about all of the complexities that have been introduced and create a little bit of test code that reads some instream data and moves it to an array. At the end of the input, display all of the entries in this array. This has nothing to do with dynamic, or comp/comp-3 data but rather how to load individual data items into an array.

Once you have the experiment working correctly, then go back to the code that gets the input from the database.

Again, there may be something i misunderstand. . . But i believe the whole topic has gone in the wrong direction for what you are stuck on.

Rijit · Active User Joined: 15 Apr 2010 Posts: 168 Location: Pune

Ronald,

You made avalid point regarding usage of 'select *'. I was aware of it and in my code I have coded all the columns insead of *. But since the number of columns was around 100 so tats why just I used * to denote it here..But good catch

dbzTHEdinosauer · Posted: Thu Sep 02, 2010 10:43 am

Roland has provided a very good example of how you avoid
wasting machine time initializing a cobol internal table

only thing I would comment on is
SET ARRAY-INDEX TO 1
MOVE ZERO TO SQLCODE
PERFORM UNTIL SQLCODE NOT = 0