Hello,
If I write tmp = CE_COLUMN_TABLE(<table>), is then tmp a pointer to <table> or does it really hold the data of <table>?.
If the latter is the case, it could happen that we get easily a memory overflow if the tables involved are sufficiently big and if the chain of CE-functions in a calculation view (CE_COLUMN_TABLE, CE_PROJECTION, CE_JOIN, ...) is long enough: Every intermediate result would be hold in its own variable. Is this really the case?
Regards,
Ingo