Well, yes, I wanted the selectivity, but I do not use PLANalyzer and
do not know what these figures indicate (I'm used to numbers between 0
and 1). I was thinking of something more along the lines of

>Index IDX_BOOKING_CLIENTNO_REQ_CO_DAT (9)
>Depth: 4, leaf buckets: 3527, nodes: 339755
>Average data length: 4.00, total dup: 167405,
>max dup: 158

(stolen from another post in this group)

I wanted these figures simply to get an impression of how common is
each value - I do not know if BOLD_TYPE is as selective as address or
"unselective" as gender. If I remember correctly, then your plan
contained two indexes on one table that was evaluated as #3 in that
part of your plan, and that made me worry a bit. Generally, it is no
disaster if the first table uses several indexes, but the further down
it gets, the greater impact it may have on performance.

Set

--- In firebird-support@yahoogroups.com, "karlzeift" wrote:
> I'm not really sure what you mean, but I think it's the
> "Selectivity" you are thinking about... is that correct?
>
> IX_ENTITY_CUSTOMER: 38.88
> IX_UNIFIEDMESSAGE_BOLD_TYPE: 4147.5
> IX_UNIFIEDMESSAGE_ENTITY: 15.74
>
> I'm using PLANalazyer to get these values.
>
> Karl
>
> --- In firebird-support@yahoogroups.com, "Svein Erling"
> <svein.erling.tysvaer@k...> wrote:
> > What are the index statistics for
> >
> > IX_ENTITY_CUSTOMER, IX_UNIFIEDMESSAGE_BOLD_TYPE and
> > IX_UNIFIEDMESSAGE_ENTITY? And approximately how many records have
> each
> > of the values specified in your where clause?
> >
> > Set