Fetch Unique and sorted data from two Cobol tables

deepak_munjal · New User Joined: 30 May 2008 Posts: 43 Location: Mumbai

Hi,

I have two input sorted COBOL tables (Array1 and Array2). I want to fetch the first column of the both the tables without duplicates and in sorted order, extracted in the first column of 3rd Cobol table(Array3).
No condition on rest of the fields of input Cobol tables.

Structure-
Array1(Emp_Id1, Amount1, Access_code1)
Array2(Emp_Id2, Amount2, Access_code2)
Array3((Emp_Id3, Name3, Amount3, Amount4, Amount5)

Emp_Id -- Pic 9(10)

Please see the below example for your reference :-

Array1 --

EMP_ID1
-------------

Anuj Dhawan · Posted: Mon Sep 17, 2012 3:09 pm

This sounds like a simple "file-match" kind of exercise - what have you tried?
Have you looked at the sticky in COBOL Forum for "2-File Match/Merge sample code". This should give you a idea to get going.

OTOH, First, COBOL has TABLEs not ARRAYs. You've used both of them but ARRAY is not a correct notion when you're in a COBOL Forum. Second, one doesn't 'fetch' values from COBOL Tables nor do they have columns.

deepak_munjal · New User Joined: 30 May 2008 Posts: 43 Location: Mumbai

Anuj,

Thanks for your response.

dbzTHEdinosauer · Posted: Mon Sep 17, 2012 3:49 pm

deepak_munjal,

yes SORT control statements (merge is a completely different thing)
to input two files and perform the matching and reporting necessary.

there is no need to CALL SORT from a COBOL program to accomplish this,
the fact that you have two COBOL Internal Tables
and want to populate a third,
just means that you have to write the code necessary
to do the same thing that a Stand-alone Sort (with the appropriate control statements)
would do.

regardless of the methodology you decide to implement for your solution,
you will have to restate your requirements.

Bill Woodger · Posted: Mon Sep 17, 2012 3:55 pm

From what has been described so far, why not put data in the third table when putting data in either the first or the second table. If duplicate in first, don't attempt. If duplicate in second, don't attempt. If already in third (previously inserted from the "other" table), don't attempt.

Cobol doesn't have a lot of "do magic" "functions", so best not to expect them in the first place. The intrinsic functions that there are are listed in the manuals, link at the top of the page.

deepak_munjal · New User Joined: 30 May 2008 Posts: 43 Location: Mumbai

Thanks Bill for your sugestion.

Anuj Dhawan · Posted: Mon Sep 17, 2012 6:57 pm

Your description about what you want to do and/or doing is confusing:

For example, when you say

dick scherrer · Posted: Mon Sep 17, 2012 10:01 pm

Hello,

enrico-sorichetti · Posted: Mon Sep 17, 2012 11:24 pm

there is one simple change ...
the standard two file match code issues ( when needed ) ONE read at the time for each file

when there are duplicates You will need to keep reading until the key is different,

does not seem overly complicated

dick scherrer · Posted: Tue Sep 18, 2012 12:22 am

Hello,

I'm still not sure i understand if there are 2 inputs or 3 . . .

Maybe TS will clarify.

deepak_munjal · New User Joined: 30 May 2008 Posts: 43 Location: Mumbai

Dick,

Bill Woodger · Posted: Tue Sep 18, 2012 3:00 am

Mmmm.... not quite what I suggested.

If you have your two "Cobol tables", which have been loaded with data "en mass", then to get your third "Cobol table" is a variation of the "two-file match", with the "Cobol tables" instead of files.

Starting with the first entry in both tables, copy the lowest key from the two tables to Table 3. Increase subscript/index for table just used - at end of one table, copy remaining entries from other table.

The only extra you need is to deal with duplicates. There are two types you have to deal with, "internal" duplicates in each of the two tables, and duplicates between the two tables.

You might find it simpler to deal with both cases at once, in the paragraph/section you write to add the entry to Table 3. If the new key for Table 3 is the same as the previous key for table 3, do not add it.

dick scherrer · Posted: Tue Sep 18, 2012 11:40 pm

Hello,

I believe this is being made more difficult than it needs to be. . .

I do not understand why so many if-else statememts are needed. I believe the only if-else code needed should be the "match" code from the sample code in the "sticky".

Keep in mind that you can ORDER both cursors by the "match keys". There should be no need for the "temp array".

The newly posted data does Not include all of the fields you mention - why not? You know what you have/want - as yet, we do not, so you need to provide it.

I still have no clue as to where the majority of the fields in Array3 come from it these are being generated by the code:
Array3((Emp_Id3, Name3, Amount3, Amount4, Amount5)

dbzTHEdinosauer · Posted: Wed Sep 19, 2012 12:24 am

have you thought of just using one CURSOR with a well written SELECT/WHERE Clause?

the problem requirement has never been well described,
but I would bet, 1 cursor and then process the FETCHed rows
and you probably will not need any COBOL Internal Tables.

This sounds like some concoction from Rookie-Central:

two db2 cursors loading into 2 COBOL Internal Tables
to load yet another COBOL Internal Table........

this is probably a CICS program.........