MVSFORUMS.com

manu · Beginner Joined: 26 Dec 2002 Posts: 47 Topics: 19

There is a idea to create name value pair tables so that it will be extensible(no need to add columns etc.) but want to understand the pros and cons of it in a Mainframe DB2 Cobol environment. Pease let me know if you have any experiences and inputs on the pros and cons. If there are other related discussions or articles, please let me know. Thanks.

kolusu · Posted: Tue Mar 14, 2023 10:33 am Post subject:

manu,

As far as I know z/OS Db2 does NOT support name value pair tables.
_________________
Kolusu - DFSORT Development Team (IBM)
DFSORT is on the Web at:
www.ibm.com/storage/dfsort

www.linkedin.com/in/kolusu

manu · Beginner Joined: 26 Dec 2002 Posts: 47 Topics: 19

sorry for the delayed response. Yes, it is not supported but it will be indirectly done by creating two columns which are generic, the first one will hold the column name and the second one will hold the value. for example

instead of
Key-Column Field1 Field2 Field3 Field4

it will be like below.
Key-Column Generic-field-name field-value
1 Field1 XXXX
2 Field2 YYYYY
3 Field3 AAAA
4 Field4 BBBB

The only advantage I see is that fields can be added with out making physical table changes. I wanted to check whether there are any other advantages in doing this pattern in mainframe DB2 Cobol environment. Please let me know if anyone had implemented something similar and their feedback. Thanks

misi01 · Posted: Fri Mar 22, 2024 2:56 am Post subject:

So, reading 4 rows instead of 1, or did I misunderstand you?

To say it sounds inefficient must be an understatement.

How many times per day will it be read?
_________________
Michael