Removing values from dictionary that multiple keys share

Question

I've got a dictionary built of title:substrings where the first three entries look like:

'1' : ['G', 'GA', 'GAA', 'GAAA', 'GAAAA', 'GAAAAA', 'GAAAAAG', 'GAAAAAGU', 'GAAAAAGUA', 'GAAAAAGUAU', 'GAAAAAGUAUG', 'GAAAAAGUAUGC', 'GAAAAAGUAUGCA', 'GAAAAAGUAUGCAA', 'GAAAAAGUAUGCAAG', 'GAAAAAGUAUGCAAGA', 'GAAAAAGUAUGCAAGAA', 'GAAAAAGUAUGCAAGAAC']

'2' : ['G', 'GA', 'GAG', 'GAGA', 'GAGAG', 'GAGAGA', 'GAGAGAG', 'GAGAGAGA', 'GAGAGAGAC', 'GAGAGAGACA', 'GAGAGAGACAU', 'GAGAGAGACAUA', 'GAGAGAGACAUAG', 'GAGAGAGACAUAGA', 'GAGAGAGACAUAGAG', 'GAGAGAGACAUAGAGG']

'3' : ['G', 'GU', 'GUC', 'GUCU', 'GUCUU', 'GUCUUU', 'GUCUUUG', 'GUCUUUGU', 'GUCUUUGU"', 'GUCUUUGU"G', 'GUCUUUGU"GU', 'GUCUUUGU"GUA', 'GUCUUUGU"GUAC', 'GUCUUUGU"GUACA', 'GUCUUUGU"GUACAU', 'GUCUUUGU"GUACAUC']

I found quite a few answers on how to convert them to lists and creating a for loop to check if items in 1 is in other and if so remove them, but I don't think that's feasible for me since this dictionary is generated and not handwritten by me. Because of that I don't really understand how I should be approaching this problem.

Answer 1

You could build an index where each value lists the keys it is in. If that list is greater than 1, remove that value.

mydict = {
'1' : ['G', 'GA', 'GAA', 'GAAA', 'GAAAA', 'GAAAAA', 'GAAAAAG', 'GAAAAAGU', 'GAAAAAGUA', 'GAAAAAGUAU', 'GAAAAAGUAUG', 'GAAAAAGUAUGC', 'GAAAAAGUAUGCA', 'GAAAAAGUAUGCAA', 'GAAAAAGUAUGCAAG', 'GAAAAAGUAUGCAAGA', 'GAAAAAGUAUGCAAGAA', 'GAAAAAGUAUGCAAGAAC'],

'2' : ['G', 'GA', 'GAG', 'GAGA', 'GAGAG', 'GAGAGA', 'GAGAGAG', 'GAGAGAGA', 'GAGAGAGAC', 'GAGAGAGACA', 'GAGAGAGACAU', 'GAGAGAGACAUA', 'GAGAGAGACAUAG', 'GAGAGAGACAUAGA', 'GAGAGAGACAUAGAG', 'GAGAGAGACAUAGAGG'],

'3' : ['G', 'GU', 'GUC', 'GUCU', 'GUCUU', 'GUCUUU', 'GUCUUUG', 'GUCUUUGU', 'GUCUUUGU"', 'GUCUUUGU"G', 'GUCUUUGU"GU', 'GUCUUUGU"GUA', 'GUCUUUGU"GUAC', 'GUCUUUGU"GUACA', 'GUCUUUGU"GUACAU', 'GUCUUUGU"GUACAUC']
}

# build a dict of values and the keys that hold them
tmpdict = defaultdict(list)
for key, values in mydict.items():
    for value in values:
        tmpdict[value].append(key)

# now remove offenders from the original list
for value, inkeys in tmpdict.items():
    if len(inkeys) > 1:
        for key in inkeys:
            mydict[key].remove(value)

print mydict