I've seen quite a lot of answers on flattening dictionaries here but couldn't really apply any of them to my case. I have an example of the structure of the dictionary below and the output that I think would be ideal but not necessary. Anything to flatten it in a way that points to the value instead of the key would be great.

(most of the answers I found here were always using the key and that wouldn't be very helpful in this situation)

    b = [
          {'name': 'Level A',
           'children': [
               {'name': 'Level B',
                'children': [
                    {'name': 'Level C'},
                    {'name': 'Level D'},
                    {'name': 'Level E',
                     'children': [
                         {'name': 'Level Z'}
                     ]}
                ]
                },
               {'name': 'Level F',
                'children': [
                    {'name': 'Level G'}
                ]}
           ]
           },
          {'name': 'Level M',
           'children': [{'name': 'Level N'},
                        {'name': 'Level Y'}
                        ]
           }
      ]

Wanted output:

l = ['Level A-Level B-Level C',
     'Level A-Level B-Level D',
     'Level A-Level B-Level E-Level Z',
     'Level A-Level F-Level G',
     'Level M-Level N',
     'Level M-Level Y']

Thank you all in advance!

CodePudding user response：

Recursion yield makes short work of this.

def flatten(b):
    for d in b:
        if 'children' in d:
            for child in flatten(d['children']):
                yield d['name']   '-'   child
        else:
            yield d['name']

result = [*flatten(b)]

result contains

['Level A-Level B-Level C',
 'Level A-Level B-Level D',
 'Level A-Level B-Level E-Level Z',
 'Level A-Level F-Level G',
 'Level M-Level N',
 'Level M-Level Y']

Note that the last two lines are different than your desired output, but I assume that's because you made an error. (Level M is not a child of Level A. It exists on the same level.)