Apparently, it's a widely held myth that a large # of extents (let's
say "BETWEEN 5 AND 1000") per table segment is bad for performance. Yet
the same sources who label the belief mistaken persist in pushing for
fitting all of a table in the INITIAL extent.
And that confuses the heck out of me.
This really depends on the type of access. If it's random using indexes, the extents
don't matter much. If you do a lot of full table scans, and the extents are scattered
all over, there may be a performance degradation. YMMV as always.
[EMAIL PROTECTED] 02/27/01 01:06PM
Apparently, it's