Nested JSON to Parquet

Consider a nested JSON having the following format:

{
    "f1": "v1",
    "f2": 1,
    "f3": [{
        "a": "b",
        "c": "d"
    }, {
        "e": "f",
        "g": "h"
    }],
    "f4": {
        "f5": "v5",
        "f6": {
            "f7": [1,2,3,4,5,6],
            "f8": "v8"
        }
    }
}

If I do not flatten it and use Glue ETL to convert it to parquet directly, would it be the correct approach? Would the storage of nested maps/arrays in parquet columns be ideal if I do it this way?