Ann,

Oops. I expected your response in IB-Architect list, but it's already here.
That's wonderful ;-)

>>I have one relatively big table (~20000 records) and a few small ones (0 to
>>1000 records) in my database (6MB). When I'm trying to select all fields
>>from the big table left joined with all others without sorting, it works
>>fine. When I'm trying to perform the same statement with ORDER BY, a
>>temporary sorting file is being created by the engine, but this file is
>>finally ~600MB in size. Simple calculation shows that the complete result
>>set for my query should have the similar size, but only in the case when
>>all fields are stored with their maximum possible size, for example ~255
>>bytes for VARCHAR(255) and so on. Is it true or it's just coincidence?
>
>No, it's not a coincidence. To avoid re-reading the database in suboptimal
>order, InterBase produces an intermediate result set and performs a tag
>sort (I believe) on those intermediate results. The intermediate set is
>stored in fixed sized buffers - thus the expansion of varchar(255) to 257
>bytes.

Thanks a lot for your explanation. But it seems to me, that in situations
when tables contain large character fields, buffer I/O will be the most
part of a query execution. This will be extremely sad when most of those
character fields are nulls or empty.

Best regards,
Dmitry