Re: Index getting "old" fast

Ok, perhaps. But if there were 1,000 rows of data in the table what are the chances that Oracle would even look at the index? My expectation would be that a full table scan would be faster than reading the index followed by reading the table.

From what he describes it appears that the index is being used ... therefore my assumption is that the table contains a lot of data. so adding a few thousand rows shouldn't make significant difference.

I'm more than happy to be wrong on this one ... but it just seems unlikely based on the following:

1 day = 2880 rows
1 year = 751,000 rows (assuming weekends and holidays off) Is another 2880 rows going to make the statistics that far wrong?

My instinct is that there is something else going on. Meaning far more transactions than what he describes; or worse.

Daniel Morgan

Niall Litchfield wrote:

> "damorgan" <dan.morgan_at_ci.seattle.wa.us> wrote in message
> news:3C72A226.BA453A6_at_ci.seattle.wa.us...
> > 2880 inserts are making your indexes go bad or requiring new statistics? I
> don't buy it.
>
> I do. 13 rows made our finance app go bad. Makes sense of course. 13 rows =
> periods for a new financial year. reanalyze worked fine. So sure in general
> an extra days worth of collected data won't make a significant difference
> but it migfht do in some situations - like there are only three days worth
> of data in the table to start with. So it all depends.
>
> either way an overnight analyze is likely to do the job adequately.
>
> --
> Niall Litchfield
> Oracle DBA
> Audit Commission UK
> *****************************************
> Please include version and platform
> and SQL where applicable
> It makes life easier and increases the
> likelihood of a good answer
>
> ******************************************
Received on Mon Feb 25 2002 - 11:08:04 CST