Clash Royale CLAN TAG

Python: DIvide items of a dictionary by a value that depends on the first key

I believe that this questions is very similar to Divide each python dictionary by total value, however I am not able to deal with multidimentional key.

So I have this first dictionary:

x={'SBGL': 0.2631678018921199, 'SBAR': 0.00017329658914466182, 'SBKP': 0.05787906885949929, 'SBRJ': 0.01686853063415596, 'SBFZ': 0.002151268003175112, 'SBCH': 0.0, 'SBRF': 0.0010995369794006128, 'SBFI': 0.0, 'SBGR': 0.3555667545433087, 'SBCT': 0.12645173241811486, 'SBFL': 0.0020536641771051302, 'SBSP': 0.12206752790423671, 'SBSV': 0.002883296698699977, 'SBMG': 0.0, 'SBCA': 0.0, 'SBPA': 0.024485612897250063, 'SBPS': 0.0006125138064595806, 'SBCF': 0.005023609170377438, 'SBMO': 0.0020686035382382908, 'SBBR': 0.01744718188871371, 'SBJV': 0.0}


I have this second dictionary:

y= {('SBCF', 'TAM'): 294.0, ('SBCH', 'GLO'): 0.0, ('SBSP', 'ONE'): 0.0, ('SBGL', 'TAM'): 114094.0, ('SBCA', 'PTB'): 0.0, ('SBKP', 'GLO'): 0.0, ('SBMG', 'AZU'): 0.0, ('SBRF', 'GLO'): 1104.0, ('SBPA', 'AZU'): 23367.0, ('SBGR', 'AZU'): 313.0, ('SBGR', 'GLO'): 105170.0, ('SBCT', 'ONE'): 0.0, ('SBKP', 'TAM'): 160.0, ('SBGL', 'ONE'): 25330.0, ('SBGR', 'ONE'): 69043.0, ('SBCT', 'PTB'): 0.0, ('SBRJ', 'TAM'): 8118.0, ('SBPS', 'GLO'): 615.0, ('SBPA', 'TAM'): 1218.0, ('SBSV', 'GLO'): 1691.0, ('SBFI', 'AZU'): 0.0, ('SBSP', 'TAM'): 62158.0, ('SBKP', 'ONE'): 0.0, ('SBGR', 'PTB'): 0.0, ('SBGR', 'TAM'): 182484.0, ('SBCT', 'AZU'): 85180.0, ('SBBR', 'TAM'): 12685.0, ('SBBR', 'GLO'): 4833.0, ('SBGL', 'GLO'): 124812.0, ('SBJV', 'GLO'): 0.0, ('SBCT', 'TAM'): 11374.0, ('SBCT', 'GLO'): 30411.0, ('SBMO', 'GLO'): 1822.0, ('SBCF', 'AZU'): 4750.0, ('SBPA', 'GLO'): 0.0, ('SBCF', 'GLO'): 0.0, ('SBMO', 'TAM'): 255.0, ('SBFL', 'AZU'): 2062.0, ('SBFL', 'GLO'): 0.0, ('SBCA', 'AZU'): 0.0, ('SBRJ', 'GLO'): 7426.0, ('SBFI', 'ONE'): 0.0, ('SBKP', 'AZU'): 57954.0, ('SBFZ', 'TAM'): 97.0, ('SBSP', 'GLO'): 60405.0, ('SBRJ', 'AZU'): 1393.0, ('SBFZ', 'GLO'): 2063.0, ('SBSV', 'TAM'): 1204.0, ('SBAR', 'TAM'): 174.0, ('SBFI', 'GLO'): 0.0}


I want to divide all the items of y by a value that depends on the first key and comes from x. To be more precise, all the items of y that has the first key equal to 'SBCF' will be divided by x['SBCF']. All the items of y that has the first key equal to has the first key equal to 'SBCH' should be divided by x['SBCH'] and so on...

I tried something like this:

y = {[i,j]: v / (x[i]) for i,j,v in y.items()}


I have got the error

not enough values to unpack (expected 3, got 2)


I want to divide all the items of y by a value that depends on the first key and comes from x.

4 Answers
4

Wrap i and j with an additional set of parentheses, since y.items() returns something of the form <(k1, k2), v>.

i

j

y.items()

<(k1, k2), v>

Some things to account for:

x

{(i, j) : v / max(x.get(i, 1), 1) for (i, j), v in y.items()}


You can either choose to handle the division by 0, or let an exception be thrown. Additionally, if i is missing in x, you should think how you'd want to handle that. I've assumed nothing is done (i.e., division by 1).

i

x

tuple

list

You can try something like this

y = {k: v / (x[k[0]]) for k,v in y.items() if x[k[0]]}


you can try:

{k: v / (x[k[0]] or 1) for k,v in y.items()}


where x[k[0]] or 1 fix ZeroDivisionError

x[k[0]] or 1

ZeroDivisionError

The error is due to the part i,j,v in y.items().
y.items() returns a enumerable collection of 2 item types of the form (<key>,<value>).
Since here your, in itself, a tuple, you can unpack it further by wrapping the i,j in brackets so: (i,j),v in y.items() which will unpack the key item pair and then unpack the compound key.

i,j,v in y.items()

y.items()

(<key>,<value>)

i,j

(i,j),v in y.items()

Overall this gives: y = {[i,j]: v / (x[i]) for (i,j),v in y.items()}

y = {[i,j]: v / (x[i]) for (i,j),v in y.items()}

You may also encounter the issue that the key cannot be a list, as lists cannot be hash, so you will have to substitute [i,j] for (i,j).

[i,j]

(i,j)

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.