On Mon, 5 Aug 2019 10:55:00 +1000, Andrew Rowley wrote:

>On 1/08/2019 8:03 am, Ron Thomas wrote:
>>
>> We have a FB File LRECL = 80 bytes and here below is the layout. Buy_price 
>> is at position 65 of length 10 bytes . We need to pull those item nbrs
>> which has same UPC,Vendor nbr ,State Code but there is price difference. 
>> Could anyone let me know how this to be done using dfsort.
>
>I am curious, why do you want to use Sort to perform these tasks?
> 
+1
I'm probably partisan, but I'd look to Rexx or awk, both of which provide
useful associative arrays.

>The solutions are ingenious, but I wonder whether someone will be able
>to understand them if modification is required in the future.
>
>Maybe a more general programming language would be better?
>
Perhaps Sort is the language most generally understood among the
readers of the list.  Or the one that elicits the most useful answers.


On Thu, 1 Aug 2019 09:21:22 -0400, Cameron Conacher wrote:
>...
>I have a file of data containing extracts from some JCL.
>Specifically, I am looking for all values of DATACLAS.
>This all relates back to Pervasive Encryption. I just want a quick report
>of the DATACLAS values in use today, as well as the counts for each
>DATACLAS item I find.
>
>I ran a quick compare to get a report from our JCL Library, and now I am
>building a SORT to look at the variable length input file data.
>...
Again, I'd look first to Rexx PARSE or sed regular expressions.  There's
considerable animus to regex in this list, but its complexity pales compared
to Sort.

I see error sources regardless:

o JCL comments might look like DD statements, causing false positives.

o Instream data sets even more so.

o JCL might be built with symbol substitution, causing false negatives.

o Likewise arguments to PROCs.

The Swiss Army Knife is a wondrous facility, but there is probably a better
tool to drive in a nail.

(OT: is DFSORT Turing-complete?)

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to