Hi,

you might have to do some recoding, if you use indexed datasets or pointer control. You might also notice that a 'compressed' dataset can be larger than an uncompressed one. This is because of the small overhead you have in the datasets for managing them. Keep in mind that less storage may cost more cpu.

This will be the case: (1) when the dataset is very small, i.e. it has very few observations (2) when the dataset has only fields that use their full length

Cheers Adolf Quast