Re: CHGPF on file with multiple format logicals.

Be sure to validate the DTLINE field definition in every format of each MFLF, to ensure it is expanded or not, according to what is the desired outcome. Do not just recompile, without that verification. If the noted error was a surprise, so too might the effects of the CHGPF for other logical files; which means validating single format DDS LF might be in order as well.

A field in an LF format is /insulated/ from picking up changes to the base field, when the LF field is explicitly named in the logical format. Only when the field being included in the LF is due to a shared format reference to the physical format, will the logical field be changed.

If in a MFLF one format shares the physical format and another does not, and if the changing field was a key field, then the shared format requires that the access path be rebuilt -- because the definition of the key field was changed. However a non-shared format in the same MFLF requires that the field attributes remain unchanged [so the access path need not be rebuilt]. It is that conflict which gives rise to the condition diagnosed by CPF3238.

To know which of the logical files have at least one format that can exhibit the updated field definition, obtain the list of files from DSPDBR FILE(thePF) RCDFMT(*ALL) and INTERSECT that data with a DSPDBR FILE(thePF) MBR(*NONE). Any MFLF in the list must be reviewed further for the condition of a changing key, to know if that will be an issue.

Regards, Chuck