Re: Possible ODBC weirdness when deleting records

Replying to myself, since I have a bit of correction/clarification:

On Wed, Mar 9, 2016 at 4:26 PM, John Yeung <gallium.arsenide@xxxxxxxxx> wrote:

Hmm... just to further explore the lock thing - when I have the
clearable file open for viewing in some other program (on the i), my
script hangs when it tries to do the delete, as though it needs to
secure an exclusive lock before proceeding. When I have the other file
open (the one which does accumulate deleted records), my script can go
ahead and delete the records, even though I'm still viewing that file
in the other program.

I tried both STRSQL and my ODBC-based script again, and when I've got
the clearable file open for reading in another program, both STRSQL
and the script will hang for a while, apparently trying to secure a
lock for the clear; but if I wait long enough, both of them will
proceed with the regular deletion. So it does appear that this file
*is* behaving as documented (SQL clears it if it can, but does
row-by-row deletion if it can't do the clear).

So, in a way, that means that all the OTHER files are the weird ones,
because SQL doesn't seem to ever even attempt to clear them, and just
goes straight to regular deletion.

Hm... have to check at least one more thing out - Charles mentioned this:

A significant number of rows are being deleted.

That was one of the required conditions for the clear to be attempted.
But these are small files that I'm looking at. The clearable one has
always had a bit more records than the others, so maybe I'm near the
boundary of what "a significant number" is. Trying to test that
next....

John Y.