The source of your data (json) has nothing to do with what you want, which is to find the dictionary in y['filters'] that contains a key called filterB. To do this, you need to iterate over the list and look for the item that fulfils this condition.
w = None
for item in y['filters']:
if 'filterB' in item:
w = item['filterB']
break
print(w) # val1
Alternatively, you could join all dictionaries into a single dictionary and use that like you originally tried
all_dict = dict()
for item in y['filters']:
all_dict.update(item)
# Replace the list of dicts with the dict
y['filters'] = all_dict
w = y['filters']['filterB']
print(w) # val1
If you have multiple dictionaries in the list that fulfil this condition and you want w to be a list of all these values, you could do:
y = {"filters":[{"filterA":"All"},{"filterB":"val1"},{"filterB":"val2"}]}
all_w = list()
for item in y['filters']:
if 'filterB' in item:
all_w.append(item['filterB'])
Or, as a list-comprehension:
all_w = [item['filterB'] for item in y['filters'] if 'filterB' in item]
print(all_w) # ['val1', 'val2']
Note that a list comprehension is just syntactic sugar for an iteration that creates a list. You aren't avoiding any looping by writing a regular loop as a list comprehension
jsonlibrary: to convert JSON documents into ordinary Python data.