Pythonic way to iterate through json file-CodePudding

I have a .json file which has a similar structure like this:

{
  "cars": [
    {
    "FIAT": [
      {"model_id": 153},
      {"model_id": 194}
    ]
    },
    {
    "AUDI": [
      {"model_id": 261}
    ]
    },
    {
    "BMW": [
      {"model_id": 264}
    ]
    }
  ]
}

My end goal is to retrieve the following:

      {"model_id": 153},
      {"model_id": 194},
      {"model_id": 261},
      {"model_id": 264}

Currently my code which gets me this result is this:

    for cars in dir['cars']:
        for brand in cars:
            for model in cars[brand]:
                print(model)

My question is that do we have a better way to access these details? I know that itertools.product is used to replace nested for loops, but can it be applied in this scenario?

CodePudding user response：

I think this is exactly the problem that ijson tries to solve, loading partial information from a json file.

CodePudding user response：

product isn't useful here, as you don't have two independent lists to compute a product from. (Nor do you actually want a product of anything). You can, however, use a list comprehension with multiple iterators

[x for c in d['cars'] for v in c.values() for x in v]

one of which can be replaced with itertools.chain operating on a generator expression:

list(chain.from_iterable(v for c in d['cars'] for v in c.values()))

You can apply chain.from_iterable again:

list(chain.from_iterable(chain.from_iterable(c.values() for c in d['cars'])))

though readability may start to suffer.

A final version that does away with all explicit iteration

from itertools import chain
from functools import methodcaller

list(chain.from_iterable(
       chain.from_iterable(
        map(methodcaller('values'), d['cars'])))

is only for the most extreme adherents to functional programming and higher-order functions :)

CodePudding user response：

You can use jq to process the json and extract what you want. Here is an example for your data https://jqplay.org/s/mIaelGNpXO