Python: Merge two dictionaries

You could start by first making a new dict using the ** operator:

new_dict = {**A, **B}
>>> new_dict
{1: 'one', 2: 'dva', 3: 'three'}

Then a for loop over the set intersection of keys:

>>> for dupe_key in set(A) & set(B):
...     new_dict[dupe_key] = [A[dupe_key], B[dupe_key]]
...
>>> new_dict
{1: 'one', 2: ['two', 'dva'], 3: 'three'}

Here's a quick solution:

A = {1 : "one", 2 : "two"}
B = {2 : "dva", 3 : "three"}
d = {**A}

for k, v in B.items():
    d[k] = [d[k], v] if k in d else v

print(d)

Output:

{1: 'one', 2: ['two', 'dva'], 3: 'three'}

Here's a more generalizable solution, that works with an indefinite number of dictionaries:

def cool_dict_merge(dicts_list):
    d = {**dicts_list[0]}
    for entry in dicts_list[1:]:
        for k, v in entry.items():
            d[k] = ([d[k], v] if k in d and type(d[k]) != list
                    else [*d[k], v] if k in d
                    else v)
    return d

Testing:

>>> A = {1: "one", 2: "two"}
>>> B = {2: "dva", 3: "three"}
>>> C = {2: "chewbacca", 3: "ThReE", 4: "four"}
>>> D = {0: "zero", 4: "jack", 5: "five"}
>>> cool_dict_merge([A, B, C, D])
{1: 'one', 2: ['two', 'dva', 'chewbacca'], 3: ['three', 'ThReE'], 4: ['four', 'jack'], 0: 'zero', 5: 'five'}

Note that, as others pointed out, the result you're currently getting is probably more preferable in most scenarios, since it's more generalizable and easier to work with. We don't know about your use case though, so just use the method you feel most comfortable with.

This code achieves what you're looking for:

A = {1 : "one", 2 : "two"}
B = {2 : "dva", 3 : "three"}

d = A.copy()
for key, val in B.items():
    if key in d:
        if isinstance(d[key], list):
            d[key].append(val)
        else:
            d[key] = [d[key], val]
    else:
        d[key] = val

print(d)

But, as mentioned in comments, it is less generalizable, meaning you have to check the type of value everytime (more if-else checks) to work with it.

I used dictionaries and groupby groups to find like sets of data. Pandas great for transforming data into dictionaries, tuples, or lists.

A = {1 : "one", 2 : "two"}
B = {2 : "dva", 3 : "three"}

z1=zip(A.keys(),A.values())
z2=zip(B.keys(),B.values())

mylist=[]
mylist.append(list(z1))
mylist.append(list(z2))

variables=[]
phrases=[]
for item in mylist:
   for k,v in item:
       variables.append(k)
       phrases.append(v)
df=pd.DataFrame(zip(variables,phrases),columns=['variable','phrase'])

grpDictionary=df.groupby('variable').groups

resultDictionary={}
for key, values in grpDictionary.items():
    resultDictionary[key]=" ".join(df.iloc[values]['phrase'])

print(resultDictionary)

output:

{1: 'one', 2: 'two dva', 3: 'three'}