Basically this was a bug in 9.1 SP2 that has been fixed in SP3. It is inconsequential except that it clogs up the tm1server.log file.This problem occurs because during the feeding process the second time the rule was changed we are changing the cell type from STET to SIMPLE. We record the change in cell type but nothing is noted about the cell data type because there was no change in the cell data; basically the cell data is expected to go back to whatever value it had before the STET rule was added. The same problem occurs if the cell was a rule calculated value and it was fed. The cell in this case would be changing from a FEEDER to SIMPLE. The solution is not trivial in this case if we want to keep the old value in the cell. It involves getting the cell value for a fed cell (which is never deleted from the data tree). Since there are a few things to think about to implement this solution and the only issues caused by this are: 1) The bogus error message that troubles users and 2) In some weird and complicated scenarios (for example a TI process that executes RuleLoadFromFile() and then CellGetN/S(), the cellget won't return the old value in the cell but 0 or an empty string)