CODEWITHSUNDEEP
pythonjson
Sylwia Malinowska
Sylwia Malinowska

Reputation: 21

Merging JSON elements with the same keys

I've been struggling with JSON transformation using Python. I have JSON in the below format: 

{
    "Children": [{ "child": "Child 0"}],
    "Parent": "Parent 10"
},
{
    "Children": [{ "child": "Child 1"}],
    "Parent": "Parent 10"
},
{
    "Children": [{ "child": "Child 2"}],
    "Parent": "Parent 11"
},

But instead of having duplicated parents, I would like to merge children together to get that:

{
    "Children": [{ "child": "Child 0"}, { "child": "Child 1"}],
    "Parent": "Parent 10"
},
{
    "Children": [{ "child": "Child 2"}],
    "Parent": "Parent 11"
},

Upvotes: 2

Views: 3772

Answers (2)

RoadRunner
RoadRunner

Reputation: 26315

You can also use a collections.defaultdict() to do this, which can be serialized at the end:

from collections import defaultdict
from json import dumps

data = [
    {"Children": [{"child": "Child 0"}], "Parent": "Parent 10"},
    {"Children": [{"child": "Child 1"}], "Parent": "Parent 10"},
    {"Children": [{"child": "Child 2"}], "Parent": "Parent 11"},
]

d = defaultdict(list)
for dic in data:
    parent, children = dic["Parent"], dic["Children"]
    d[parent].extend(children)

result = []
for k, v in d.items():
    result.append({"Parent": k, "Children": v})

print(dumps(result))

Which gives a JSON array of JSON objects:

[{"Parent": "Parent 10", "Children": [{"child": "Child 0"}, {"child": "Child 1"}]}, {"Parent": "Parent 11", "Children": [{"child": "Child 2"}]}]

You can also group into the data by parent key using a nested defaultdict():

d = defaultdict(lambda : defaultdict(list))
for dic in data:
    parent, children = dic["Parent"], dic["Children"]
    d[parent]["Children"].extend(children)

print(dumps(d))

Which gives this new structure:

{"Parent 10": {"Children": [{"child": "Child 0"}, {"child": "Child 1"}]}, "Parent 11": {"Children": [{"child": "Child 2"}]}}

And will allow easy O(1) lookups for the parent.

Upvotes: 2

Benjamin Rowell
Benjamin Rowell

Reputation: 1411

Take a look at the itertools groupby function. Here's an example with your data grouped by Parent.

>>> from itertools import groupby
>>> import pprint

>>> data = [{
    "Children": [{ "child": "Child 0"}],
    "Parent": "Parent 10"
},
{
    "Children": [{ "child": "Child 1"}],
    "Parent": "Parent 10"
},
{
    "Children": [{ "child": "Child 2"}],
    "Parent": "Parent 11"
}]

>>> data_grouped = {k: list(v) for k, v in groupby(data, key=lambda x: x["Parent"])}

>>> pp = pprint.PrettyPrinter(indent=4)
>>> pp.pprint(data_grouped)

{   'Parent 10': [   {   'Children': [{'child': 'Child 0'}],
                         'Parent': 'Parent 10'},
                     {   'Children': [{'child': 'Child 1'}],
                         'Parent': 'Parent 10'}],
    'Parent 11': [{'Children': [{'child': 'Child 2'}], 'Parent': 'Parent 11'}]}

Here I've placed your example dicts inside a list and group by the Parent entry in each dict. This is all wrapped up inside a dict comprehension to give a meaningful output.

Upvotes: 2

Related Questions