There is a JSON like this:

{
  "P1": "ss",
  "Id": 1234,
  "P2": {
      "P1": "cccc"
  },
  "P3": [
      {
          "P1": "aaa"
      }
  ]
}

How can I find all P1‘s value without it iterating all JSON?

P.S.: P1 can be anywhere in the JSON.

If no method can do this, can you tell me how to iterate through the JSON?

As I said in my other answer, I don’t think there is a way of finding all values associated with the "P1" key without iterating over the whole structure. However I’ve come up with even better way to do that which came to me while looking at @Mike Brennan’s answer to another JSON-related question How to get string objects instead of Unicode from JSON?

The basic idea is to use the object_hook parameter that json.loads() accepts just to watch what is being decoded and check for the sought-after value.

Note: This will only work if the representation is of a JSON object (i.e. something enclosed in curly braces {}), as in your sample.

from __future__ import print_function
import json

def find_values(id, json_repr):
    results = []

    def _decode_dict(a_dict):
        try:
            results.append(a_dict[id])
        except KeyError:
            pass
        return a_dict

    json.loads(json_repr, object_hook=_decode_dict) # Return value ignored.
    return results

json_repr="{"P1": "ss", "Id": 1234, "P2": {"P1": "cccc"}, "P3": [{"P1": "aaa"}]}"
print(find_values('P1', json_repr))

(Python 3) output:

['cccc', 'aaa', 'ss']

I had the same issue just the other day. I wound up just searching through the entire object and accounted for both lists and dicts. The following snippets allows you to search for the first occurrence of a multiple keys.

import json

def deep_search(needles, haystack):
    found = {}
    if type(needles) != type([]):
        needles = [needles]

    if type(haystack) == type(dict()):
        for needle in needles:
            if needle in haystack.keys():
                found[needle] = haystack[needle]
            elif len(haystack.keys()) > 0:
                for key in haystack.keys():
                    result = deep_search(needle, haystack[key])
                    if result:
                        for k, v in result.items():
                            found[k] = v
    elif type(haystack) == type([]):
        for node in haystack:
            result = deep_search(needles, node)
            if result:
                for k, v in result.items():
                    found[k] = v
    return found

deep_search(["P1", "P3"], json.loads(json_string))

It returns a dict with the keys being the keys searched for. Haystack is expected to be a Python object already, so you have to do json.loads before passing it to deep_search.

Any comments for optimization are welcomed!

My approach to this problem would be different.

As JSON doesn’t allow depth first search, so convert the json to a Python Object, feed it to an XML decoder and then extract the Node you are intending to search

from xml.dom.minidom import parseString
import json        
def bar(somejson, key):
    def val(node):
        # Searches for the next Element Node containing Value
        e = node.nextSibling
        while e and e.nodeType != e.ELEMENT_NODE:
            e = e.nextSibling
        return (e.getElementsByTagName('string')[0].firstChild.nodeValue if e 
                else None)
    # parse the JSON as XML
    foo_dom = parseString(xmlrpclib.dumps((json.loads(somejson),)))
    # and then search all the name tags which are P1's
    # and use the val user function to get the value
    return [val(node) for node in foo_dom.getElementsByTagName('name') 
            if node.firstChild.nodeValue in key]

bar(foo, 'P1')
[u'cccc', u'aaa', u'ss']
bar(foo, ('P1','P2'))
[u'cccc', u'cccc', u'aaa', u'ss']

Using json to convert the json to Python objects and then going through recursively works best. This example does include going through lists.

import json
def get_all(myjson, key):
    if type(myjson) == str:
        myjson = json.loads(myjson)
    if type(myjson) is dict:
        for jsonkey in myjson:
            if type(myjson[jsonkey]) in (list, dict):
                get_all(myjson[jsonkey], key)
            elif jsonkey == key:
                print myjson[jsonkey]
    elif type(myjson) is list:
        for item in myjson:
            if type(item) in (list, dict):
                get_all(item, key)

Converting the JSON to Python and recursively searching is by far the easiest:

def findall(v, k):
  if type(v) == type({}):
     for k1 in v:
         if k1 == k:
            print v[k1]
         findall(v[k1], k)

findall(json.loads(a), 'P1')

(where a is the string)

The example code ignores arrays. Adding that is left as an exercise.

Bearing in mind that json is simply a string, using regular expressions with look-ahead and look-behind can accomplish this task very quickly.

Typically, the json would have been extracted from a request to external api, so code to show how that would work has been included but commented out.

import re
#import requests
#import json

#r1 = requests.get( ... url to some api ...)
#JSON = str(json.loads(r1.text))
JSON = """
 {
  "P1": "ss",
  "Id": 1234,
  "P2": {
      "P1": "cccc"
  },
  "P3": [
     {
          "P1": "aaa"
     }
  ]
 }
"""
rex1  = re.compile('(?<=\"P1\": \")[a-zA-Z_\- ]+(?=\")')
rex2 = rex1.findall(JSON)  
print(rex2)

#['ss', 'cccc', 'aaa']

I don’t think there’s any way of finding all values associated with P1 without iterating over the whole structure. Here’s a recursive way to do it that first deserializes the JSON object into an equivalent Python object. To simplify things most of the work is done via a recursive private nested function.

import json

try:
    STRING_TYPE = basestring
except NameError:
    STRING_TYPE = str  # Python 3

def find_values(id, obj):
    results = []

    def _find_values(id, obj):
        try:
            for key, value in obj.items():  # dict?
                if key == id:
                    results.append(value)
                elif not isinstance(value, STRING_TYPE):
                    _find_values(id, value)
        except AttributeError:
            pass

        try:
            for item in obj:  # iterable?
                if not isinstance(item, STRING_TYPE):
                    _find_values(id, item)
        except TypeError:
            pass

    if not isinstance(obj, STRING_TYPE):
        _find_values(id, obj)
    return results

json_repr="{"P1": "ss", "Id": 1234, "P2": {"P1": "cccc"}, "P3": [{"P1": "aaa"}]}"

obj = json.loads(json_repr)
print(find_values('P1', obj))

You could also use a generator to search the object after json.load().

Code example from my answer here: https://stackoverflow.com/a/39016088/5250939

def item_generator(json_input, lookup_key):
    if isinstance(json_input, dict):
        for k, v in json_input.iteritems():
            if k == lookup_key:
                yield v
            else:
                for child_val in item_generator(v, lookup_key):
                    yield child_val
    elif isinstance(json_input, list):
        for item in json_input:
            for item_val in item_generator(item, lookup_key):
                yield item_val