Re: Index query - monotonically increasing columns

Afternoon Richard, Niall,

as ever, thanks for your comments. As you say, in this experiment, it makes not a jot of difference. My experimenting showed that with a random letter at the start, the index depth is quite happily 2 for ages and ages and ages !

In a hypothetical situation though, with millions of rows being generated on a sequence, would this be a good way of avoiding index problems - and I'm talking theory/good practice here. Pure ascending numbers being indexed with rows being deleted is going to trash the index over time. Slapping a random letter in front is going to spread the data out - similar to indexeing on someones surname perhaps - and will allow index entries to be re-used over time.

Or, is there another way .......? :o)

Cheers,
Norman.

--
Still behind the firewall, still bolloxed, and having to post from
Google.
And I'm having a crap day as well.