Re: CHGPF and "Pointer not set for location referenced." error

[CRPence <CRPbottle@xxxxxxxxx>]

  Suggesting an error like msgMCH3601 occurs without any supporting 
details has as much worth as asking why your car will not start without 
any supporting details.

  A /machine/ exception is always issued to a program [and module and 
procedure], either from a LIC program or as part of an error within the 
program itself.  In the latter case, the names are manifest the same, 
both as from\to, such that the /from/ information is omitted.  In the 
former case the error might be from program dbdoesstuff and to program 
qdbchgfi module dostuff procedure doit.  Only knowing the full context 
of that information can any worthwhile comment be made about the error 
that was received; i.e. what was all the _do_ stuff doing, as can be 
inferred by the program, module, and procedure names.  Additionally 
anything specific about the database file network and attributes of the 
PF [such as triggers, identity, blob, etc.] that might be unique for the 
failing file as compared to other files that do not exhibit the same 
failure would add supporting details which might help to identify either 
a circumvention or the origin of the problem.

  Aside from an error with the db file, the db cross reference, or the 
dbf network, the MCH3601 in an IBM program is almost surely a defect 
condition, or would be found to be a /problem with the file [network]/ 
when reported to your service provider.  If not a defect, or if the 
result of an old defect which is merely being manifest visibly in this 
manner, the recovery is typically to delete and recreate the file, using 
whatever method of data recovery is appropriate for the individual 
situation.

Regards, Chuck

Mike Cunningham wrote:
> I was doing some database work last night and ran CHGPF against 6 PFs
> successfully but then I ran it on one file it was not happy and
> issued a pair of "Pointer not set for location referenced." errors
> (MCH3601), no other lower level errors. The CHGPF command failed on a
> development version of this file and the production version. I
> removed the new field I was adding from the source and got the same
> error. Found a hit for PFT SI23243 which we already have applied and
> SI31599 which we did not so I applied that and still got the error.
> Has anyone seen this error before and have a workaround?
>
> Running V5R4