Re: Splitting a big file into smaller files

On 17-Apr-2015 15:55 -0500, Darryl Freinkel wrote:

I am splitting large files of 400,000 plus records into files of
10,000 each.

I am using SQL to create the records using the create table
command. I split the file using

WHERE RRN(A) between 00001 and 10000
WHERE RRN(A) between 10001 and 20000
WHERE RRN(A) between 20001 and 30000

Problem:
Why would the system not copy exactly 10,000 records. I have
several files that have a few records less than the 10,000 I
should have in each file.

Has anyone seen this problem before?

<http://archive.midrange.com/midrange-l/201502/msg00622.html>

If you have, what is the reason?

<http://archive.midrange.com/midrange-l/201502/msg00637.html>

This is a V5R4 system.

If the issue is the same as alluded in the thread <http://archive.midrange.com/midrange-l/201502/threads.html#00619>, in which those prior included links to archived messages were posted, then presumably the issue being encountered is independent of release and a natural consequence of the scripted actions that were performed [but for which the script was not offered in the OP] prior to the SELECT of the data by RRN.

Quite likely the issue is a natural consequence [a side effect] of both active and inactive records claiming a physical Relative Record Number (RRN) and the Reuse Delete Records (REUSEDLT) feature enabling essentially the random placement of inserted rows as active records. Any assumption that either the SQL DELETE statement assuredly effects the same as a Clear Physical File Member (CLRPFM) or that deleted rows are not potentially interspersed and thus are skipped in RRN(), are very likely to be proven to be false assumption. If the assumptions fail, then the CREATE TABLE ... SELECT ... WHERE RRN() BETWEEN requests would be expected to probably effect uneven /splitting/ of the data across those smaller files [as uneven distributions of the active-row data from the bigger file].