Hello,

ODBC or ADO will outperform your custom code in performance for sure.

Both are designed to be efficient for even large datasets.

You can't program a faster data retrival method yourself (at least not in VFP, a scripting language is slow by definition), taking into consideration additional (D)COM overhead .. your own coded marshaling code will definitly perform poor when transfering large datasets ..

if you run your posted code in a loop to retrieve a whole cursor it will be slower by a !factor! of at least 10 compared to ADO or ODBC.

Regards
Christian