CODEWITHSUNDEEP
pythonpython-2.7
GPiter
GPiter

Reputation: 799

Group a multidimensional array by a key in python

I have this code : 

 [
    {
        "info": {
            "a": 65535,
            "b": 50,
        },
        "test": {
            "m": "10300",
            "id": "2008",
        },
    },
    {
        "info": {
            "a": 65535,
            "b": 50,
        },
        "test": {
            "m": "10300",
            "id": "2008",
        },
    },
    {
        "info": {
            "a": 65535,
            "b": 50,
        },
        "test": {
            "m": "10300",
            "id": "2009",
        },
    },
]

I want to have : 

[
    "2008" : [
        {
            "info": {
                "a": 65535,
                "b": 50,
             },
            "test": {
                "m": "10300",
                "id": "2008",
             },
        },
        {
            "info": {
                "a": 65535,
                "b": 50,
            },
            "test": {
                "m": "10300",
                "id": "2008",
            },
       },
    ]
    "2009" : [
        {
            "info": {
                "a": 65535,
                "b": 50,
            },
            "test": {
                "m": "10300",
                "id": "2009",
            },
        },
     ]

The idea is to group by key value based on column id. I'm new in python an have no idea how to do that. Please help me. Thx in advance. If you could provide some ideas will be great.I found some examples but can't solve this problem anyway.

My solution : 

value = [i for i in array]
res = sorted(value, key=lambda x: x["id"], reverse=True)

Upvotes: 0

Views: 5925

Answers (3)

KBA7
KBA7

Reputation: 41

Let us suppose

array = [
   {
    "info": {
        "a": 65535,
        "b": 50,
        },
    "test": {
        "m": "10300",
        "id": "2008",
        },
    },
    {
    "info": {
        "a": 65535,
        "b": 50,
        },
    "test": {
        "m": "10300",
        "id": "2008",
        },
    },
    {
    "info": {
        "a": 65535,
        "b": 50,
        },
    "test": {
        "m": "10300",
        "id": "2009",
        },
    },
]

Then you can get the expected result by using 

modified_data = {}
for entry in array:
    index = entry['test']['id']
    try:
        modified_data[index].append(entry)
    except KeyError:
        modified_data[index] = [entry]
print modified_data

Now let's analyze your code

value = [i for i in array]

After this value will be equal to array itself, assuming that your array initialization is same as mine. 

res = sorted(value, key=lambda x: x["id"], reverse=True)

If the above statement is correct then value will not have 'id' hence it will throw an error.

Also, a list can't have a key-value pair. You have to use a dictionary for that.

Upvotes: 1

Ozzy Walsh
Ozzy Walsh

Reputation: 887

I think you want a dictionary where each key is the id and the value is your object.

Try the following. data is your array of data.

new = {}
for item in data:
    item_id = item['test']['id']
    new[item_id] = item

You could also use a dictionary comprehension.

new = {item['test']['id']: item for item in data}

Upvotes: 3

Sunitha
Sunitha

Reputation: 12015

Use itertools.groupby

from itertools import groupby
f = lambda d: d['test']['id']
res = {k:list(v) for k,v in groupby(sorted(l, key=f), f)}
pprint(res)

Output

{'2008': [{'info': {'a': 65535, 'b': 50},
           'test': {'id': '2008', 'm': '10300'}},
          {'info': {'a': 65535, 'b': 50},
           'test': {'id': '2008', 'm': '10300'}}],
 '2009': [{'info': {'a': 65535, 'b': 50},
           'test': {'id': '2009', 'm': '10300'}}]}

Upvotes: 1

Related Questions