MVSFORUMS.com

shiv_swami · Beginner Joined: 29 Nov 2003 Posts: 68 Topics: 14

Hi All,

I have a following requirement in which I need your help to decide on what tool/language to use.
I have extract data from IDMS Database everyday & send the files in an incremental way.

If file has key of 10 chars & remaining data of 2400 CHARACTERS

YESTERDAY EXTRACT

kolusu · Posted: Thu Jan 15, 2004 11:43 am Post subject:

Shiv,

Just change the definition for WS-TOD-DESC & WS-YEST-DESC. Specify the absolute start position for the compare fields which is 11 instead of letting easytrieve calculate it.
The reason is simple. You have defined the enitre RECORD with WS-TODAY-REC and next fields are just a redefinition of the ws-today-rec. When you use '*' for specifying field positions you cannot use redefinition.

so you can either change it to

shiv_swami · Beginner Joined: 29 Nov 2003 Posts: 68 Topics: 14

Kolusu,

Thank you very much for the solution Laughing

.

The new definition of the variables is correct & I am getting the desired results.
I would like to know whether there would be any performance issues due to high record length (upto 3000) & high number of records in the Matched files (appx 10000 to 15000).

Also I need to do this for 12 different file types..Is this the best solution ?
What are the factors that I have to consider before implementing this solution Question

Regards,
Shivprakash
Why Walk when you can Run

?
Nicola Iacocca

kolusu · Posted: Fri Jan 16, 2004 6:35 am Post subject:

shiv,

I don't think performance will be an issue. why don't you just run a couple of tests and compare the cpu times.

shiv_swami · Beginner Joined: 29 Nov 2003 Posts: 68 Topics: 14

Kolusu,
Thanks for your response.
As per your suggestion,I will do test for the large record length program.I was happy with CPU time it used for small number of records in match files.
For the 12 files ,that I want to create incremental files, they do not have same record layout.They have different layouts.(Sorry I did not make my requirements clear straight away)

kolusu · Posted: Fri Jan 16, 2004 9:21 am Post subject:

Shiv,

The following is a general easytrieve program which can be used for any file length up to 9999. I assumed that your key of 10 bytes always starts in position 1. A brief explanation of the program.

Unlike cobol , easytrieve allows you to define the files as workareas. so we define both the input files as workarea and define only the key fields on both the files.

Define another working storage variable W-TODAY-REC which is redefined as an array of 1 byte character field with an occurs clause of 9999. The same applies for the definition of
W-YESTDAY-REC

Now using reference modification we move the contents of the file to another working storage variable which are used as comparision fields.

But for reference modification to work we need to know the lrecl of the file. For this we use the variable RECORD-LENGTH which is a two-byte binary field used for all file types to determine or establish the length of the current data record. For variable-length records, this field contains only the length of the record's data.

we use a START proc(as it execuetes only once at the begining) to find LRECL of both the files . once we have the LRECL of the files we calculate the no: of bytes of data to be compared by subtracting the key field length from the TOTAL LRECL.

shiv_swami · Beginner Joined: 29 Nov 2003 Posts: 68 Topics: 14

Thanks Kolusu.

It works well & with definition of outfile as FB (0 0) ,with this one need not bother about the record lenth of output fille too as it will be on JCL card.
Key-length is the only thing I have to bother about.... Laughing

Regards,
Shivprakash
Why Walk when you can Run

?
Nicola Iacocca