Functional dependencies aren't usually expressed in a diagram. Instead, they're usually expressed like this . . .
A->B
B->CD
or like this.
"Salesperson number" -> "Salesperson name"
"Client number" -> "Client name"
If your diagram means that "Product number" -> "Quantity sold", then your diagram is wrong. Also, client number does not determine product number.
A functional dependency answers the question, "Given one value for A, can I determine one and only one value for B?", where A and B are sets of attributes. Let's look at a simple example.
Given one value for "Product number", can we determine one and only one value for "Product description"? Yes. Given the product number P1023-01, we can determine only one value for "Product description": "A4 Paper". Product number P1023-01 never matches up with any product name but "A4 Paper".
Given one value for "Product number", can we determine one and only one value for "Quantity sold"? No. Given the product number P1023-01, we can determine two different values for "Quantity sold": 6 and 10. The dependency "Product number" -> "Quantity sold" doesn't hold in this relation.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…