Rick Debay wrote:

> Would it be worthwhile to submit a feature request that these utilities
> circumvent the page cache, or is this not feasible due to architecture
> or implementation?

It's not feasible to circumvent the page cache. It _might_ be possible
(not easy) to limit the number of pages a particular query can use, or
identify a massive serial read and restrict it to a few data pages at
a time. It will still need pointer pages, etc.

Regards,


Ann