Skip to content

Instantly share code, notes, and snippets.

@douglasmiranda
Last active January 18, 2024 05:33
Show Gist options
  • Save douglasmiranda/5127251 to your computer and use it in GitHub Desktop.
Save douglasmiranda/5127251 to your computer and use it in GitHub Desktop.
# This is a really old post, in the comments (and stackoverflow too) you'll find better solutions.
def find(key, dictionary):
for k, v in dictionary.iteritems():
if k == key:
yield v
elif isinstance(v, dict):
for result in find(key, v):
yield result
elif isinstance(v, list):
for d in v:
for result in find(key, d):
yield result
example = {'app_url': '', 'models': [{'perms': {'add': True, 'change': True, 'delete': True}, 'add_url': '/admin/cms/news/add/', 'admin_url': '/admin/cms/news/', 'name': ''}], 'has_module_perms': True, 'name': u'CMS'}
list(find('admin_url', example))
@russellballestrini
Copy link

russellballestrini commented Nov 10, 2015

I took this code, modified it a bit to:

@daxonwax
Copy link

beautiful. thankyou.

@robfrei
Copy link

robfrei commented May 28, 2016

Very nice. Thanks!

@abdullahaftab
Copy link

Perfect !

@Aspen129
Copy link

This is exactly what I needed to deal with nested dict's from an API request; thank you!!

@smopucilowski
Copy link

for item in iterator: yield item can be replaced with yield from iterator.

@takenoto
Copy link

Many thanks 👍 !!!

@ajay-manwani
Copy link

Thanks a Lot !!

@truongphan
Copy link

thanks you so much

@alon-unifi
Copy link

alon-unifi commented Apr 2, 2018

Nice! if any of your nested lists contains primitives (e.g, strings) and not always dicts, it will error out. In that case add a check for it and skip the element (added line 2 before last):

def find(key, dictionary):
    for k, v in dictionary.iteritems():
        if k == key:
            yield v
        elif isinstance(v, dict):
            for result in find(key, v):
                yield result
        elif isinstance(v, list):
            for d in v:
                if isinstance(d, dict):
                    for result in find(key, d):
                        yield result

@jenweber
Copy link

Python3 users should replace dictionary.iteritems() with dictionary.items(). iteritems was removed.

Thanks for sharing!

@dannyblaker
Copy link

Awesome thanks

@Maxhirez
Copy link

Nice work-just what I needed!

@treblekicker
Copy link

Beautiful, thank you!

@abhiram24
Copy link

Super sweet and clean. Thank you!

For Python 3.x users, just replace " for k, v in dictionary.iteritems():" with " for k, v in dictionary.items():"

@PatrikHlobil
Copy link

PatrikHlobil commented Sep 24, 2018

Thanks for your nice example. I changed it a bit for my use case, where I needed to get all contained keys of a nested iterable as a list (also for the values) as can be seen in https://gist.github.com/PatrikHlobil/9d045e43fe44df2d5fd8b570f9fd78cc :

def iterate_all(iterable, returned="key"):
    
    """Returns an iterator that returns all keys or values
       of a (nested) iterable.
       
       Arguments:
           - iterable: <list> or <dictionary>
           - returned: <string> "key" or "value"
           
       Returns:
           - <iterator>
    """
  
    if isinstance(iterable, dict):
        for key, value in iterable.items():
            if returned == "key":
                yield key
            elif returned == "value":
                if not (isinstance(value, dict) or isinstance(value, list)):
                    yield value
            else:
                raise ValueError("'returned' keyword only accepts 'key' or 'value'.")
            for ret in iterate_all_keys(value, returned=returned):
                yield ret
    elif isinstance(iterable, list):
        for el in iterable:
            for ret in iterate_all_keys(el, returned=returned):
                yield ret

For the example:

example = {'app_url': '', 'models': [{'perms': {'add': True, 'change': True, 'delete': True}, 
           'add_url': '/admin/cms/news/add/', 'admin_url': '/admin/cms/news/', 'name': ''}], 
           'has_module_perms': True, 'name': u'CMS'}

It gives the following output:

list(iterate_all(example, "key"))
>>> ['app_url', 'models', 'perms', 'add', 'change', 'delete', 'add_url', 'admin_url', 'name', 'has_module_perms', 'name']

list(iterate_all(example, "value"))
>>>['', True, True, True, '/admin/cms/news/add/', '/admin/cms/news/', '', True, 'CMS']

@agustinscigliano
Copy link

Is there a way to respect the hierarchy? So for example, if i have:

[
    {
        "id": 800,
        "children": [
            {
                "id": 801,
                "children": [
                    {
                        "id": 804,
                    },
                    {
                        "id": 805,
                     }
    },
]

I get: 800/801/804 , 800/801/805 (and every combination possible)

@rameshrvr
Copy link

@PatrikHlobil You can use the module nested_lookup for fetching the keys(https://pypi.python.org/pypi/nested-lookup).
Example:

>>> example = {'app_url': '', 'models': [{'perms': {'add': True, 'change': True, 'delete': True}, 
...            'add_url': '/admin/cms/news/add/', 'admin_url': '/admin/cms/news/', 'name': ''}], 
...            'has_module_perms': True, 'name': u'CMS'}
>>> from nested_lookup import get_all_keys
>>> get_all_keys(example)
['app_url', 'models', 'perms', 'add', 'change', 'delete', 'add_url', 'admin_url', 'name', 'has_module_perms', 'name']
>>> 

@praveenbandi86
Copy link

Can you pl give me the explanation of the code.

@vaibhavQ
Copy link

Hey, can you give the code for deleting all the entries in dict matching the key we send.

@Sunil-ghodela
Copy link

Thanks...

@AvdN
Copy link

AvdN commented Apr 10, 2019

This is broken, except for the limited example you give where you have a dict at the root of your data-structure and restrictions on lists.
If you just add another key to your example:

example['xyz'] = [[dict(admin_url="broken")]]

You don't get two results, as one would expect, you get an attributeerror.

That StackOverflow post has much better answers, you should not indiscriminately copy the one with the unrealistic assumptions. You should also indicate this is Python2 only because of your use of .iteritems()

@Palash90
Copy link

Thank you very much, this saved my day.

@futoase
Copy link

futoase commented Dec 29, 2019

Thanks! 🙏

@vaibhavk69
Copy link

i have this produced this raw data and now i only want latitude and longitude pairs, so can anybody help me...........

[11:40 PM, 1/12/2020] Vaibhav Kaushik: {"Response": {"MetaInfo": {"Timestamp": "2020-01-12T18:02:12.167+0000"}, "View": [{"Result": [{"Location": {"Address": {"AdditionalData": [{"key": "CountryName", "value": "India"}, {"key": "StateName", "value": "Uttar Pradesh"}, {"key": "CountyName", "value": "Gautam Buddha Nagar"}], "City": "Noida", "Country": "IND", "County": "Gautam Buddha Nagar", "District": "Sector 125", "Label": "Amity Road, Sector 125, Noida 201303, India", "PostalCode": "201303", "State": "UP", "Street": "Amity Road"}, "DisplayPosition": {"Latitude": 28.54712, "Longitude": 77.33482}, "LocationId": "NT_8RywhVafbhYYOKQDacPkWA", "LocationType": "point", "MapView": {"BottomRight": {"Latitude": 28.54405, "Longitude": 77.33731}, "TopLeft": {"Latitude": 28.5493, "Longitude": 77.33237}}, "NavigationPosition": [{"Latitude": 28.54712, "Longitude": 77.33482}]}, "MatchLevel": "street", "MatchQuality": {"City": 1.0, "Country": 1.0, "District": 1.0, "State": 1.0, "Street": [1.0]}, "Relevance": 1.0}], "ViewId": 0, "_type": "SearchResultsViewType"}]}}

@Palash90
Copy link

Hey Vaibhav, you can use the following -

def find(key, dictionary):
    for k, v in dictionary.items():
        if k == key:
            yield dictionary
        elif isinstance(v, dict):
            for result in find(key, v):
                yield result
        elif isinstance(v, list):
            for d in v:
                if isinstance(d, dict):
                    for result in find(key, d):
                        yield result

print(list(find("Latitude", data)))

I used it on your data and it is showing the following result, hope this helps you -

[{'Latitude': 28.54712, 'Longitude': 77.33482}, {'Latitude': 28.54405, 'Longitude': 77.33731},
        {'Latitude': 28.5493, 'Longitude': 77.33237}, {'Latitude': 28.54712, 'Longitude': 77.33482}]

@chiranjeevijp
Copy link

{'destination_addresses': ['Chennai, Tamil Nadu 600009, India'], 'origin_addresses': ['Chennai, Tamil Nadu 600101, India'], 'rows': [{'elements': [{'distance': {'text': '7.7 mi', 'value': 12327}, 'duration': {'text': '38 mins', 'value': 2285}, 'status': 'OK'}]}], 'status': 'OK'}

Hi Friends, from this JSON i need to extract the value '7.7' which is under the key 'text' under 'distance'. Can any one help in this

@chiranjeevijp
Copy link

chiranjeevijp commented May 6, 2020 via email

@beeoss
Copy link

beeoss commented Apr 8, 2021

Hello all.. Could anyone help me Python parse and print < "name" , "raw" > (values only), from the JSON below?

{"count":57,"next":null,"previous":null,"results":[{"id":229,"state":"complete","substate":null,"exceptions":[],"name":"Sender Account Number","output_name":null,"field_definition_attributes":{"required":false,"data_type":"Account Number","multiline":false,"routing":false,"supervision_override":null},"transcription":{"raw":"1957-3549-2","normalized":"195735492","source":"machine_transcription","data_deleted":false,"user_transcribed":null},"field_image_url":"/api/v5/image/613cf762-b4bc-46f0-a511-3dc8bb37eae3?start_x=0.3110429607297866&start_y=0.1052441592299208&end_x=0.5696909842243418&end_y=0.16043316955780607","page_id":5}

So for the 'Item' included above, starting at "results": and nesting node to "transcription", the parsed values need to be:

Sender Account Number, 1957-3549-2
(repeating items) 'for'


Much thanks for any assistance!!

@sahil-developer
Copy link

sahil-developer commented Jul 27, 2021

Thanks for the help.
Python3 users should replace dictionary.iteritems() with dictionary.items(). iteritems was removed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment