Using the above input dataset, my sort card should be
SORT FIELDS=COPY
INCLUDE COND=(15,1,CH,EQ,C'A',OR,
15,1,CH,EQ,C'B',OR,
15,1,CH,EQ,C'C')
Number of record and value A, B and C will vary in the input file. So that i need to write sort card in runtime.
We need to keep the above said job in daily run proc, so when the job running for the next day, the previous day sort card should be cleaned and need to update newly, by using that days input dataset values.
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
Hello,
You need to provide some kind of clarification. . .
The "sort card" posted shows nothing but literals
Post a more representative set of sample data, the output you want from that input, and any "rules" for selecting some part of the input to the output.
Here after I am going to get a huge volume of data in one dataset.
using that, client is expecting the following,(as the volume of data is huge, client dont want to do two file matching):
1) First we need to extract unique company names from the user provided dataset (Eg A, B, C in my sort card), so we did this one by using simple sort card (sum fields = none). Daily the list of company will vary.
2) In second step, We need to write rexx step,based on the above (Step 1) to generate a sort card.
3) Using the sort card generated in step 2, we will retrieve data from master file in step 3. Like that the process is going....
Joined: 17 Oct 2006 Posts: 2481 Location: @my desk
Quote:
Here after I am going to get a huge volume of data in one dataset
Which sort product are your using? What if it's gonna create a huge sort card so that your sort product is unable to process the control card itself?.
Quote:
as the volume of data is huge, client dont want to do two file matching
Is n't this a bit strange. If you're using SYNCSORT, you can do this in 2 steps without having to create sort cards dynamically.
Step1. As mentioned by you
Step2. Match both the files using a SyncSort JOIN application.
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
Hello,
Quote:
as the volume of data is huge, client dont want to do two file matching
Maybe i don't need to understand this, and (so far) i surely do not
If the data in 2 files is in the same key sequence, it is as fast (possibly faster) to match the 2 files as it is to pass the file, select a huge amount, and then do some processing on the extracted data.
Your request mentions generating a sort card, but it is not clear what the entire process will be. If we better understood the process, we might be able to offer more usable suggestions.
As Arun mentions, there is also the consideration of too many unique companies.