Before you read the rest, I found at least one other related posted here http://dynamicsuser.net/forums/p/45267/229112.aspx
I am going to try and explain it with as much detail as I can. . .
Dynamics AX 4.0 (Kernel 4.0.2501.116, Application 4.0.2501.121)
Let the item be called PROITEM, it’s inventory model is FIFO. It’s dimension group consists of warehouse, location and serial number and the warehouse is the primary stocking dimension. Physical inventory is ticked for all three and financial inventory for the warehouse alone.
The problem was first noticed when the InventSum table was viewed in detail, then we found these “ghost” rows with a financial value but without physical quantity. The explanation given was related to the dimension group setup, where we only have warehouse ticked for financial inventory. This made sense and we adjusted our reporting practices accordingly. Considering cost only at the warehouse level of the item. Still, this would seem to only affect those items that are modelled on an average cost principal using standard cost model.
Anyway, the above does not or more like did not bother us until this happened. The following are the inventory transactions for only the affected serial number, there are other transactions for this item, for other serial numbers.
*Transaction *Physical Date *Financial Date *Quantity *Warehouse *Cost posted *Cost adjusted
*Purchase *May 2009 *November '09 *01 *A *13,333.60 *0
*Sales *May 2009 *August '09 *-01 *A *13,333.60 *0
*Sales *July 2010 *July 2010 *01 *B *12,174.40 *0
*Sales *Sept. 2011 *Sept. 2011 *-01 *B *-6,087.20 *6087.20
Now this particular serial number was first sold, then returned almost a year later and credited back into another warehouse. Both warehouse are setup as type “default” but neither of them are the main warehouse. On the third transaction, the credit, we have a return cost of 12,174.40 - this would be equal to the inventory module table, module inventory cost price.
The problem transaction is that of September 2011, from looking at inventory transactions, it is clear that the module inventory cost price of the item was 12,068.40 so it should have been -12,068.40 at least. But the last sales transaction shows -6,087.20. Then there is the adjustment of the same amount to make the final cost of the transaction 0.
Now, we run inventory recalculation every week, and close every 6 months. We also do a end of month recalculation where we recalculate the transaction up to the end of the previous ledger period and then stop the period immediately afterwards. The last recalculation for this was 7 days after the financial date. And this is what posted the equal and opposite adjustment.
Now here is what I don’t understand. We have a copy of the company in our TEST environment and it is from January 2011. On this copy, PROITEM shows that there is 1x physical quantity with a posted value of 12,174.40 in warehouse B.
Since the PROD environment does not show any other transaction between January and September, it is a fair assumption that the financial cost was same.
That said, what the last inventory recalculation should have done was adjust the -6,087.20 (regardless of where that came from) by posting and adjustment of -6087.20, instead it posted 6087.20 and the resulting transaction was of 0 cost.
What it also did was leave behind 12,117.40 as a “ghost” transaction without any quantity.
So, what is the problem? Does inventory recalculation have a bug? Should it have posted a negative quantity but instead it posted a positive value?
But why is there a value left behind without quantity? Surely, this value should be removed on financial posting?
Shouldn’t the original cost should be -12,117.40 instead of -6087.20 and the adjustment should be 0?