问题
I created a function that accepts multiple arguments of dictionaries, and returns a concatenated dictionary. I researched online for a while about concatenating a merging dictionaries and tested the interesting ones. They all resulted in updating the values (or overwriting them).
My use case is passing in dictionaries where each key has a single value, and want a dictionary with the same or different keys, with a list of values for each key. That is my definition of what a so-called "concatenation" of dictionaries would look like.
Here are two very basic dictionaries:
a = {1: 'a', 2: 'b', 3: 'c'}
b = {1: 'd', 2: 'e', 3: 'f'}
Here is the function:
def merge_dict(*args:dict):
result = {}
for arg in args:
if not isinstance(arg, dict):
return {}
result_keys = result.keys()
for key, value in arg.items():
if key not in result_keys:
result[key] = [value]
else:
result[key].append(value)
return result
The output is:
print(merge_dict(a, b))
{1: ['a', 'd'], 2: ['b', 'e'], 3: ['c', 'f']}
I could do the same for tuples, or arrays, Numpy arrays, etc. Note this function is very simple and doesn't sanitize input or validate the data structure further than it being a dict
instance.
But, I would like to know if there is a more efficient or "pythonic" way of doing this. Please feel free to add your input.
Consider adding these dictionaries with different keys:
c = {4: 'g', 5: 'h', 6: 'i'}
d = {4: 'j', 5: 'k', 6: 'l'}
The output is:
print(merge_dict(a, b, c, d))
{1: ['a', 'd'], 2: ['b', 'e'], 3: ['c', 'f'], 4: ['g', 'j'], 5: ['h', 'k'], 6: ['i', 'l']}
I will work on nested data structures soon.
Because of your answers, here is what I did:
import collections
def merge_dicts_1(*args):
rtn = collections.defaultdict(list)
for input_dict in args:
for key, value in input_dict.items():
rtn[key].append(value)
return rtn
def merge_dicts_2(*args):
rtn = {}
for input_dict in args:
for key, value in input_dict.items():
rtn.setdefault(key, []).append(value)
return rtn
if __name__ == "__main__":
a = {1: 'a', 2: 'b', 3: 'c'}
b = {1: 'd', 2: 'e', 3: 'f'}
c = {4: 'g', 5: 'h', 6: 'i'}
d = {4: 'j', 5: 'k', 6: 'l'}
e = merge_dicts_1(a, b, c, d)
f = merge_dicts_2(a, b, c, d)
print(e)
print(f)
print(e == f)
This prints the following:
defaultdict(<class 'list'>, {1: ['a', 'd'], 2: ['b', 'e'], 3: ['c', 'f'], 4: ['g', 'j'], 5: ['h', 'k'], 6: ['i', 'l']})
{1: ['a', 'd'], 2: ['b', 'e'], 3: ['c', 'f'], 4: ['g', 'j'], 5: ['h', 'k'], 6: ['i', 'l']}
True
Thank you!
回答1:
Something like this would work for any number of input dictionaries:
import collections
def merge_dicts(*args):
rtn = collections.defaultdict(list)
for input_dict in args:
for key, value in input_dict.items():
rtn[key].append(value)
return rtn
The trick is using the defaultdict structure to automatically make new entries when they don't exist. In this case, accessing a key that doesn't yet exist creates it as an empty list.
Note that the above returns a defaultdict
object. If this isn't desirable, you can cast it back to dict or use this function instead:
def merge_dicts(*args):
rtn = {}
for input_dict in args:
for key, value in input_dict.items():
rtn.setdefault(key, []).append(value)
return rtn
回答2:
How about something like this?
from functools import reduce
def _merge_two_dicts(combined, dictionary):
for key, value in dictionary.items():
combined.setdefault(key, []).append(value)
return combined
def merge_dicts(*dicts):
return reduce(_merge_two_dicts, dicts, {})
if __name__ == '__main__':
a = {1: 'a', 2: 'b', 3: 'c'}
b = {1: 'd', 2: 'e', 3: 'f', 4: 'g'}
c = {1: 'h', 3: 'i', 5: 'j'}
combined = merge_dicts(a, b, c)
print(combined)
Output:
{1: ['a', 'd', 'h'], 2: ['b', 'e'], 3: ['c', 'f', 'i'], 4: ['g'], 5: ['j']}
来源:https://stackoverflow.com/questions/53109960/how-best-to-merge-the-values-from-multiple-dictionaries