CellIncrementN and BatchUpdate Mode

[paulsimon]

Hi

I have just found an issue whereby CellIncrementN when you are in BatchUpdate Mode works like CellPutN

This is present in at least versions 10.2 IF1 and 10.2 FP1 IF1.

This has been reported to IBM.

Regards

Paul Simon

[Duncan P]

I'm not sure that I would have accepted that as a bug if I had been IBM, given that the stated behaviour of BatchUpdateMode is that values written to cells are not available to the TI until the transaction is committed, so you would always be incrementing the value that was there at the start of the transaction.

Is that in fact what is happening? If the TI starts with a value in the cell, is the value at the end the last value applied or is it the sum of the original value and the last value applied?

[mvaspal]

I just wanted to report this bug to the forum when I saw Paul had already done so, I ran into this just today.

Why I would think this is a bug is that batch update mode works fine with the old logic:
vOldValue = CellGetN(...
CellPutN(Value + vOldValue , ...

I would expect the CellIncrementN should do the same.

[Gabor]

I thought it is more or less the TI version of the deferred update in the GUI, a kind of ancestor of sandboxing giving a private version of a cube structure without consolidations & rules. N-element values should be accessible at that time.