Re: Why non-keyed PFs back then? Why today? ( was Re: Which of the SYSIBM tables/views show the row count for )

Tom Liotta wrote:

Peter Dow (ML) wrote:

For what it's worth, these days at least, you can read a keyed
PF in arrival sequence with RPG by leaving the "K" off of the
F-spec position 34 (Record Address Type).

You could 'back then' too. The program would just fail when it hit the block(s) that were damaged. In the worst cases, that meant at the time that the first block was accessed.

It wasn't a matter of 'K' or no 'K'. It was whether or not data management would provide buffers with records without sending an *ESCAPE message.

Hmmmm. OK... Reading this with my prior message might be required for understanding.

This reply of 'hit the block(s) that were damaged' confuses the issue. That is because the rows that are not accessible by sequential access would indicate data [i.e. dataspace] damage. That is a different animal than index damage.

The topic had redirected to the key; i.e. index [aka access path] damage. My prior response was directed toward index damage only... which as Peter alludes can be avoided by using arrival sequence. But if the dataspace is damaged also, then /direct/ access method or at least RRN positioning prior to continued arrival access must be used to bypass any inaccessible rows.

Regards, Chuck