Hi

*dancing* got the book, got the book :)

but back to the subject as the book raised me sharply when I read
about string data types.

I've always thought that defining the field as varchar is a way to
avoid large memory consumption.

I have a table that store text data, and although the data itself
rarely exceed 50 characters I set it to be varchar (1024).

The select often returns hundreds of records and so my question is:
do I use 1 MB of client station memory to return 1000 records even
if they contain on average 50 characters? or do I use 50 kB?
Does server internally use 100 MB to perform grouping from 100.000
records?

That is quite important to me, I'd have to redesign entire database
to acomodate blobs instead of varchars in such case.

CUIN Kaczy