>For example check Thread #1011797
>It's a sample where SQL is faster.

At first glance I'd also say, that sql is better
for doing such mass data processing.

But I'm missing the xbase variant of that update
to really proof that sql is faster here.

The initial though was, that an sql update should
be faster than a scan loop. It was never thought of
the power of SET KEY TO to scan through the portion
of the data of a month or an agent with the appropriate
index, which could even win over the sql update.

Bye, Olaf.