How do I use dictionary keys as an index in a loop and as a value?

Question

I'm in Python and trying to produce an output from cycling through a dictionary created after running an algorithm. The dictionary represents unmet demand for network k, node i, product g, at time t.

I have a loop that indexes correctly and shows me the sum total across network k, node i, and product g.

timep = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]
# dnode is a tuplelist with three keys: (k,i,g)
objvaluetime = {}
for t in timep:
         objvaluetime[t] = sum(mydict_weightn[(k, i,g)] * (mydict_phite[(k, i,g)] - solutiondnode[k, i, g, t]) for k,i,g in dnode)/ sum(mydict_weightn[(k,i,g)] * mydict_phite[(k, i,g)] for k, i,g in dnode)

Output:

{1: 0.05890098869573218,
 2: 0.05890098869573218,
 3: 0.020467694724039757,
 4: 0.020467694724039757,
 5: 0.020467694724039757,
 6: 0.020467694724039757,
 7: 0.020467694724039757,
 8: 0.020467694724039757,
 9: 0.020467694724039757,
 10: 0.020467694724039757,
 11: 0.020467694724039757,
 12: 0.020467694724039757,
 13: 0.020467694724039757,
 14: 0.020467694724039757,
 15: 0.020467694724039757,
 16: 0.020467694724039757,
 17: 0.020467694724039757,
 18: 0.020467694724039757,
 19: 0.020467694724039757,
 20: 0.020467694724039757}

Now I'd like to do something similar, except further breaking down the unmet demand at time t = 1-20 by product g (which takes on a value 1-11).

I'd like the output to represent the values for (1,1), (1,2), (1,3) . . . (1,11), (2,1), (2,2) . . .(20,11)

This is my loop so far, but it is not correctly indexing by product g, it is simply showing the same value for all g which is incorrect.

objvaluetimecommodity = {} 
for t in timep:
    for k,i,g in dnode:
        objvaluetimecommodity[t,g] = sum(mydict_weightn[(k, i,g)] * (mydict_phite[(k, i,g)] - solutiondnode[k, i, g, t]) for k,i,g in dnode)/ sum(mydict_weightn[(k,i,g)] * mydict_phite[(k, i,g)] for k, i,g in dnode)

Output:

{(1, 1): 0.05890098869573218,
 (1, 2): 0.05890098869573218,
 (1, 3): 0.05890098869573218,
 (1, 4): 0.05890098869573218,
 (1, 5): 0.05890098869573218,
 (1, 6): 0.05890098869573218,
 (1, 7): 0.05890098869573218,
 (1, 8): 0.05890098869573218,
 (1, 9): 0.05890098869573218,
 (1, 10): 0.05890098869573218,
 (1, 11): 0.05890098869573218,
 (2, 1): 0.05890098869573218 . . .}

Edit: Essentially, I want to do what was done in objvaluetime and break it down further by product g. The values in objvaluetime represent unmet demand at a particular time. I want to see unmet demand by product (g), which happens to be one of the keys, so the sum expression may need to be modified.

Answer 1

(Answer for clarity, not a real "answer")

Your variables k,i,g in your for loop:

for k,i,g in dnode:

are being shadowed by your generator expression:

sum(
    mydict_weightn[(k,i,g)] * (
        mydict_phite[(k,i,g)] - solutiondnode[k, i, g, t]
    ) for k,i,g in dnode                                # <- k,i,g used again here
)

It's really hard to tell from your description how you should solve this.

If you want to clarify what you want we might be able to help more.