Re: Converting large amount of data

As Mark notes, it has nothing to do with record format names, but with the record format itself.

I shouldn't have stuck the word name in there; the IBM document doesn't use that word.  I think I did it because I don't use the same record name for two different database record formats.  I only use the record format from the physical when the logical has all the fields from the physical.  I know you don't have to do that, but that's the way I've always done it.

Then again, in some installations I have used the same name in all the display files (RECORD1, SFILE1, FKEYS and so on), so there you go.

On 1/28/2021 5:23 PM, Mark Waterbury wrote:

John,

A very nicely constructed test scenario.  :-)

The record format level ID for LF2 no longer matches because the total record length of LF2 is now different than the record lengths for the PF or for LF1.  The hashing algorithm that computes the level ID considers the record length, and the types, lengths, and decimal places of each component field in the record, from left to right. Note that it does not take into consideration the field names, or the field text descriptions, etc.

Record format sharing is also a rather complex subject that has never been well-documented or well explained, except perhaps in some internals documentation within the walls of IBM RTC (aka. RochLabs).


Mark

On Thursday, January 28, 2021, 5:57:27 PM EST, smith5646midrange@xxxxxxxxx <smith5646midrange@xxxxxxxxx> wrote:

It doesn't matter if the format names match in the PF and LFs.

I created the following DDS members
JRSTESTPF
R RTESTPF
  FIELD1        10
  FIELD2        10
  FIELD3        10
K FIELD1
K FIELD2

JRSTESTLF1
R RTESTPF                  PFILE(JRSTESTPF)
K FIELD2
K FIELD1

JRSTESTLF2
R RTESTPF                  PFILE(JRSTESTPF)
  FIELD1        10
  FIELD2        10
  FIELD3        10
K FIELD3
K FIELD2

Note that JRSTESTLF2 has the fields included in the DDS.

The record format IDs on all three files match.

I changed FIELD1 from 10 to 20
JRSTESTPF    R RTESTPF
  FIELD1        20
  FIELD2        10
  FIELD3        10
K FIELD1
K FIELD2

And then do a CHGPF on JRSTESTPF
FIELD1 in JRSTESTPF is changed to 20
FIELD1 in JRSTESTLF1 is changed to 20
FIELD1 in JRSTESTLF2 is NOT changed.  It remains 10 bytes.

Also, the record format ID on JRSTESTPF and JRSTESTLF1 match.  JRSTESTLF2 no longer matches.