Develop Reference

RuntimeError: dictionary changed size during iteration using Python3? [duplicate]

I have a dictionary of lists in which some of the values are empty:
d = {'a': [1], 'b': [1, 2], 'c': [], 'd':[]}
At the end of creating these lists, I want to remove these empty lists before returning my dictionary. I tried doing it like this:
for i in d:
if not d[i]:
d.pop(i)
but I got a RuntimeError. I am aware that you cannot add/remove elements in a dictionary while iterating through it...what would be a way around this then?
See Modifying a Python dict while iterating over it for citations that this can cause problems, and why.

In Python 3.x and 2.x you can use use list to force a copy of the keys to be made:
for i in list(d):
In Python 2.x calling keys made a copy of the keys that you could iterate over while modifying the dict:
for i in d.keys():
But note that in Python 3.x this second method doesn't help with your error because keys returns an a view object instead of copying the keys into a list.

You only need to use copy:
This way you iterate over the original dictionary fields and on the fly can change the desired dict d.
It works on each Python version, so it's more clear.
In [1]: d = {'a': [1], 'b': [1, 2], 'c': [], 'd':[]}
In [2]: for i in d.copy():
...: if not d[i]:
...: d.pop(i)
...:
In [3]: d
Out[3]: {'a': [1], 'b': [1, 2]}
(BTW - Generally to iterate over copy of your data structure, instead of using .copy for dictionaries or slicing [:] for lists, you can use import copy -> copy.copy (for shallow copy which is equivalent to copy that is supported by dictionaries or slicing [:] that is supported by lists) or copy.deepcopy on your data structure.)

Just use dictionary comprehension to copy the relevant items into a new dict:
>>> d
{'a': [1], 'c': [], 'b': [1, 2], 'd': []}
>>> d = {k: v for k, v in d.items() if v}
>>> d
{'a': [1], 'b': [1, 2]}
For this in Python 2:
>>> d
{'a': [1], 'c': [], 'b': [1, 2], 'd': []}
>>> d = {k: v for k, v in d.iteritems() if v}
>>> d
{'a': [1], 'b': [1, 2]}

This worked for me:
d = {1: 'a', 2: '', 3: 'b', 4: '', 5: '', 6: 'c'}
for key, value in list(d.items()):
if value == '':
del d[key]
print(d)
# {1: 'a', 3: 'b', 6: 'c'}
Casting the dictionary items to list creates a list of its items, so you can iterate over it and avoid the RuntimeError.

I would try to avoid inserting empty lists in the first place, but, would generally use:
d = {k: v for k,v in d.iteritems() if v} # re-bind to non-empty
If prior to 2.7:
d = dict( (k, v) for k,v in d.iteritems() if v )
or just:
empty_key_vals = list(k for k in k,v in d.iteritems() if v)
for k in empty_key_vals:
del[k]

To avoid "dictionary changed size during iteration error".
For example: "when you try to delete some key",
Just use 'list' with '.items()'. Here is a simple example:
my_dict = {
'k1':1,
'k2':2,
'k3':3,
'k4':4
}
print(my_dict)
for key, val in list(my_dict.items()):
if val == 2 or val == 4:
my_dict.pop(key)
print(my_dict)
Output:
{'k1': 1, 'k2': 2, 'k3': 3, 'k4': 4}
{'k1': 1, 'k3': 3}
This is just an example. Change it based on your case/requirements.

For Python 3:
{k:v for k,v in d.items() if v}

You cannot iterate through a dictionary while it’s changing during a for loop. Make a casting to list and iterate over that list. It works for me.
for key in list(d):
if not d[key]:
d.pop(key)

Python 3 does not allow deletion while iterating (using the for loop above) a dictionary. There are various alternatives to do it; one simple way is to change the line
for i in x.keys():
with
for i in list(x)

The reason for the runtime error is that you cannot iterate through a data structure while its structure is changing during iteration.
One way to achieve what you are looking for is to use a list to append the keys you want to remove and then use the pop function on dictionary to remove the identified key while iterating through the list.
d = {'a': [1], 'b': [1, 2], 'c': [], 'd':[]}
pop_list = []
for i in d:
if not d[i]:
pop_list.append(i)
for x in pop_list:
d.pop(x)
print (d)

For situations like this, I like to make a deep copy and loop through that copy while modifying the original dict.
If the lookup field is within a list, you can enumerate in the for loop of the list and then specify the position as the index to access the field in the original dict.

Nested null values
Let's say we have a dictionary with nested keys, some of which are null values:
dicti = {
"k0_l0":{
"k0_l1": {
"k0_l2": {
"k0_0":None,
"k1_1":1,
"k2_2":2.2
}
},
"k1_l1":None,
"k2_l1":"not none",
"k3_l1":[]
},
"k1_l0":"l0"
}
Then we can remove the null values using this function:
def pop_nested_nulls(dicti):
for k in list(dicti):
if isinstance(dicti[k], dict):
dicti[k] = pop_nested_nulls(dicti[k])
elif not dicti[k]:
dicti.pop(k)
return dicti
Output for pop_nested_nulls(dicti)
{'k0_l0': {'k0_l1': {'k0_l2': {'k1_1': 1,
'k2_2': 2.2}},
'k2_l1': 'not '
'none'},
'k1_l0': 'l0'}

The Python "RuntimeError: dictionary changed size during iteration" occurs when we change the size of a dictionary when iterating over it.
To solve the error, use the copy() method to create a shallow copy of the dictionary that you can iterate over, e.g., my_dict.copy().
my_dict = {'a': 1, 'b': 2, 'c': 3}
for key in my_dict.copy():
print(key)
if key == 'b':
del my_dict[key]
print(my_dict) # 👉️ {'a': 1, 'c': 3}
You can also convert the keys of the dictionary to a list and iterate over the list of keys.
my_dict = {'a': 1, 'b': 2, 'c': 3}
for key in list(my_dict.keys()):
print(key)
if key == 'b':
del my_dict[key]
print(my_dict) # 👉️ {'a': 1, 'c': 3}

If the values in the dictionary were unique too, then I used this solution:
keyToBeDeleted = None
for k, v in mydict.items():
if(v == match):
keyToBeDeleted = k
break
mydict.pop(keyToBeDeleted, None)

Deleting dictionary keys from a provided list in Python [duplicate]

I know how to remove an entry, 'key' from my dictionary d, safely. You do:
if d.has_key('key'):
del d['key']
However, I need to remove multiple entries from a dictionary safely. I was thinking of defining the entries in a tuple as I will need to do this more than once.
entities_to_remove = ('a', 'b', 'c')
for x in entities_to_remove:
if x in d:
del d[x]
However, I was wondering if there is a smarter way to do this?

Using dict.pop:
d = {'some': 'data'}
entries_to_remove = ('any', 'iterable')
for k in entries_to_remove:
d.pop(k, None)

Using Dict Comprehensions
final_dict = {key: value for key, value in d if key not in [key1, key2]}
where key1 and key2 are to be removed.
In the example below, keys "b" and "c" are to be removed & it's kept in a keys list.
>>> a
{'a': 1, 'c': 3, 'b': 2, 'd': 4}
>>> keys = ["b", "c"]
>>> print {key: a[key] for key in a if key not in keys}
{'a': 1, 'd': 4}
>>>

Why not like this:
entries = ('a', 'b', 'c')
the_dict = {'b': 'foo'}
def entries_to_remove(entries, the_dict):
for key in entries:
if key in the_dict:
del the_dict[key]
A more compact version was provided by mattbornski using dict.pop()

a solution is using map and filter functions
python 2
d={"a":1,"b":2,"c":3}
l=("a","b","d")
map(d.__delitem__, filter(d.__contains__,l))
print(d)
python 3
d={"a":1,"b":2,"c":3}
l=("a","b","d")
list(map(d.__delitem__, filter(d.__contains__,l)))
print(d)
you get:
{'c': 3}

If you also need to retrieve the values for the keys you are removing, this would be a pretty good way to do it:
values_removed = [d.pop(k, None) for k in entities_to_remove]
You could of course still do this just for the removal of the keys from d, but you would be unnecessarily creating the list of values with the list comprehension. It is also a little unclear to use a list comprehension just for the function's side effect.

Found a solution with pop and map
d = {'a': 'valueA', 'b': 'valueB', 'c': 'valueC', 'd': 'valueD'}
keys = ['a', 'b', 'c']
list(map(d.pop, keys))
print(d)
The output of this:
{'d': 'valueD'}
I have answered this question so late just because I think it will help in the future if anyone searches the same. And this might help.
Update
The above code will throw an error if a key does not exist in the dict.
DICTIONARY = {'a': 'valueA', 'b': 'valueB', 'c': 'valueC', 'd': 'valueD'}
keys = ['a', 'l', 'c']
def remove_key(key):
DICTIONARY.pop(key, None)
list(map(remove_key, keys))
print(DICTIONARY)
output:
DICTIONARY = {'b': 'valueB', 'd': 'valueD'}

Some timing tests for cpython 3 shows that a simple for loop is the fastest way, and it's quite readable. Adding in a function doesn't cause much overhead either:
timeit results (10k iterations):
all(x.pop(v) for v in r) # 0.85
all(map(x.pop, r)) # 0.60
list(map(x.pop, r)) # 0.70
all(map(x.__delitem__, r)) # 0.44
del_all(x, r) # 0.40
<inline for loop>(x, r) # 0.35
def del_all(mapping, to_remove):
"""Remove list of elements from mapping."""
for key in to_remove:
del mapping[key]
For small iterations, doing that 'inline' was a bit faster, because of the overhead of the function call. But del_all is lint-safe, reusable, and faster than all the python comprehension and mapping constructs.

I have no problem with any of the existing answers, but I was surprised to not find this solution:
keys_to_remove = ['a', 'b', 'c']
my_dict = {k: v for k, v in zip("a b c d e f g".split(' '), [0, 1, 2, 3, 4, 5, 6])}
for k in keys_to_remove:
try:
del my_dict[k]
except KeyError:
pass
assert my_dict == {'d': 3, 'e': 4, 'f': 5, 'g': 6}
Note: I stumbled across this question coming from here. And my answer is related to this answer.

I have tested the performance of three methods:
# Method 1: `del`
for key in remove_keys:
if key in d:
del d[key]
# Method 2: `pop()`
for key in remove_keys:
d.pop(key, None)
# Method 3: comprehension
{key: v for key, v in d.items() if key not in remove_keys}
Here are the results of 1M iterations:
del: 2.03s 2.0 ns/iter (100%)
pop(): 2.38s 2.4 ns/iter (117%)
comprehension: 4.11s 4.1 ns/iter (202%)
So both del and pop() are the fastest. Comprehensions are 2x slower.
But anyway, we speak nanoseconds here :) Dicts in Python are ridiculously fast.

Why not:
entriestoremove = (2,5,1)
for e in entriestoremove:
if d.has_key(e):
del d[e]
I don't know what you mean by "smarter way". Surely there are other ways, maybe with dictionary comprehensions:
entriestoremove = (2,5,1)
newdict = {x for x in d if x not in entriestoremove}

inline
import functools
#: not key(c) in d
d = {"a": "avalue", "b": "bvalue", "d": "dvalue"}
entitiesToREmove = ('a', 'b', 'c')
#: python2
map(lambda x: functools.partial(d.pop, x, None)(), entitiesToREmove)
#: python3
list(map(lambda x: functools.partial(d.pop, x, None)(), entitiesToREmove))
print(d)
# output: {'d': 'dvalue'}

I think using the fact that the keys can be treated as a set is the nicest way if you're on python 3:
def remove_keys(d, keys):
to_remove = set(keys)
filtered_keys = d.keys() - to_remove
filtered_values = map(d.get, filtered_keys)
return dict(zip(filtered_keys, filtered_values))
Example:
>>> remove_keys({'k1': 1, 'k3': 3}, ['k1', 'k2'])
{'k3': 3}

It would be nice to have full support for set methods for dictionaries (and not the unholy mess we're getting with Python 3.9) so that you could simply "remove" a set of keys. However, as long as that's not the case, and you have a large dictionary with potentially a large number of keys to remove, you might want to know about the performance. So, I've created some code that creates something large enough for meaningful comparisons: a 100,000 x 1000 matrix, so 10,000,00 items in total.
from itertools import product
from time import perf_counter
# make a complete worksheet 100000 * 1000
start = perf_counter()
prod = product(range(1, 100000), range(1, 1000))
cells = {(x,y):x for x,y in prod}
print(len(cells))
print(f"Create time {perf_counter()-start:.2f}s")
clock = perf_counter()
# remove everything above row 50,000
keys = product(range(50000, 100000), range(1, 100))
# for x,y in keys:
# del cells[x, y]
for n in map(cells.pop, keys):
pass
print(len(cells))
stop = perf_counter()
print(f"Removal time {stop-clock:.2f}s")
10 million items or more is not unusual in some settings. Comparing the two methods on my local machine I see a slight improvement when using map and pop, presumably because of fewer function calls, but both take around 2.5s on my machine. But this pales in comparison to the time required to create the dictionary in the first place (55s), or including checks within the loop. If this is likely then its best to create a set that is a intersection of the dictionary keys and your filter:
keys = cells.keys() & keys
In summary: del is already heavily optimised, so don't worry about using it.

Another map() way to remove list of keys from dictionary
and avoid raising KeyError exception
dic = {
'key1': 1,
'key2': 2,
'key3': 3,
'key4': 4,
'key5': 5,
}
keys_to_remove = ['key_not_exist', 'key1', 'key2', 'key3']
k = list(map(dic.pop, keys_to_remove, keys_to_remove))
print('k=', k)
print('dic after = \n', dic)
**this will produce output**
k= ['key_not_exist', 1, 2, 3]
dic after = {'key4': 4, 'key5': 5}
Duplicate keys_to_remove is artificial, it needs to supply defaults values for dict.pop() function.
You can add here any array with len_ = len(key_to_remove)
For example
dic = {
'key1': 1,
'key2': 2,
'key3': 3,
'key4': 4,
'key5': 5,
}
keys_to_remove = ['key_not_exist', 'key1', 'key2', 'key3']
k = list(map(dic.pop, keys_to_remove, np.zeros(len(keys_to_remove))))
print('k=', k)
print('dic after = ', dic)
** will produce output **
k= [0.0, 1, 2, 3]
dic after = {'key4': 4, 'key5': 5}

def delete_keys_from_dict(dictionary, keys):
"""
Deletes the unwanted keys in the dictionary
:param dictionary: dict
:param keys: list of keys
:return: dict (modified)
"""
from collections.abc import MutableMapping
keys_set = set(keys)
modified_dict = {}
for key, value in dictionary.items():
if key not in keys_set:
if isinstance(value, list):
modified_dict[key] = list()
for x in value:
if isinstance(x, MutableMapping):
modified_dict[key].append(delete_keys_from_dict(x, keys_set))
else:
modified_dict[key].append(x)
elif isinstance(value, MutableMapping):
modified_dict[key] = delete_keys_from_dict(value, keys_set)
else:
modified_dict[key] = value
return modified_dict
_d = {'a': 1245, 'b': 1234325, 'c': {'a': 1245, 'b': 1234325}, 'd': 98765,
'e': [{'a': 1245, 'b': 1234325},
{'a': 1245, 'b': 1234325},
{'t': 767}]}
_output = delete_keys_from_dict(_d, ['a', 'b'])
_expected = {'c': {}, 'd': 98765, 'e': [{}, {}, {'t': 767}]}
print(_expected)
print(_output)

I'm late to this discussion but for anyone else. A solution may be to create a list of keys as such.
k = ['a','b','c','d']
Then use pop() in a list comprehension, or for loop, to iterate over the keys and pop one at a time as such.
new_dictionary = [dictionary.pop(x, 'n/a') for x in k]
The 'n/a' is in case the key does not exist, a default value needs to be returned.

Comparing values in a dictionary and grouping them again in Python [duplicate]

Given a dictionary like so:
my_map = {'a': 1, 'b': 2}
How can one invert this map to get:
inv_map = {1: 'a', 2: 'b'}

Python 3+:
inv_map = {v: k for k, v in my_map.items()}
Python 2:
inv_map = {v: k for k, v in my_map.iteritems()}

Assuming that the values in the dict are unique:
Python 3:
dict((v, k) for k, v in my_map.items())
Python 2:
dict((v, k) for k, v in my_map.iteritems())

If the values in my_map aren't unique:
Python 3:
inv_map = {}
for k, v in my_map.items():
inv_map[v] = inv_map.get(v, []) + [k]
Python 2:
inv_map = {}
for k, v in my_map.iteritems():
inv_map[v] = inv_map.get(v, []) + [k]

To do this while preserving the type of your mapping (assuming that it is a dict or a dict subclass):
def inverse_mapping(f):
return f.__class__(map(reversed, f.items()))

Try this:
inv_map = dict(zip(my_map.values(), my_map.keys()))
(Note that the Python docs on dictionary views explicitly guarantee that .keys() and .values() have their elements in the same order, which allows the approach above to work.)
Alternatively:
inv_map = dict((my_map[k], k) for k in my_map)
or using python 3.0's dict comprehensions
inv_map = {my_map[k] : k for k in my_map}

Another, more functional, way:
my_map = { 'a': 1, 'b':2 }
dict(map(reversed, my_map.items()))

We can also reverse a dictionary with duplicate keys using defaultdict:
from collections import Counter, defaultdict
def invert_dict(d):
d_inv = defaultdict(list)
for k, v in d.items():
d_inv[v].append(k)
return d_inv
text = 'aaa bbb ccc ddd aaa bbb ccc aaa'
c = Counter(text.split()) # Counter({'aaa': 3, 'bbb': 2, 'ccc': 2, 'ddd': 1})
dict(invert_dict(c)) # {1: ['ddd'], 2: ['bbb', 'ccc'], 3: ['aaa']}
See here:
This technique is simpler and faster than an equivalent technique using dict.setdefault().

This expands upon the answer by Robert, applying to when the values in the dict aren't unique.
class ReversibleDict(dict):
# Ref: https://stackoverflow.com/a/13057382/
def reversed(self):
"""
Return a reversed dict, with common values in the original dict
grouped into a list in the returned dict.
Example:
>>> d = ReversibleDict({'a': 3, 'c': 2, 'b': 2, 'e': 3, 'd': 1, 'f': 2})
>>> d.reversed()
{1: ['d'], 2: ['c', 'b', 'f'], 3: ['a', 'e']}
"""
revdict = {}
for k, v in self.items():
revdict.setdefault(v, []).append(k)
return revdict
The implementation is limited in that you cannot use reversed twice and get the original back. It is not symmetric as such. It is tested with Python 2.6. Here is a use case of how I am using to print the resultant dict.
If you'd rather use a set than a list, and there could exist unordered applications for which this makes sense, instead of setdefault(v, []).append(k), use setdefault(v, set()).add(k).

Combination of list and dictionary comprehension. Can handle duplicate keys
{v:[i for i in d.keys() if d[i] == v ] for k,v in d.items()}

A case where the dictionary values is a set. Like:
some_dict = {"1":{"a","b","c"},
"2":{"d","e","f"},
"3":{"g","h","i"}}
The inverse would like:
some_dict = {vi: k for k, v in some_dict.items() for vi in v}
The output is like this:
{'c': '1',
'b': '1',
'a': '1',
'f': '2',
'd': '2',
'e': '2',
'g': '3',
'h': '3',
'i': '3'}

For instance, you have the following dictionary:
my_dict = {'a': 'fire', 'b': 'ice', 'c': 'fire', 'd': 'water'}
And you wanna get it in such an inverted form:
inverted_dict = {'fire': ['a', 'c'], 'ice': ['b'], 'water': ['d']}
First Solution. For inverting key-value pairs in your dictionary use a for-loop approach:
# Use this code to invert dictionaries that have non-unique values
inverted_dict = dict()
for key, value in my_dict.items():
inverted_dict.setdefault(value, list()).append(key)
Second Solution. Use a dictionary comprehension approach for inversion:
# Use this code to invert dictionaries that have unique values
inverted_dict = {value: key for key, value in my_dict.items()}
Third Solution. Use reverting the inversion approach (relies on the second solution):
# Use this code to invert dictionaries that have lists of values
my_dict = {value: key for key in inverted_dict for value in my_map[key]}

Lot of answers but didn't find anything clean in case we are talking about a dictionary with non-unique values.
A solution would be:
from collections import defaultdict
inv_map = defaultdict(list)
for k, v in my_map.items():
inv_map[v].append(k)
Example:
If initial dict my_map = {'c': 1, 'd': 5, 'a': 5, 'b': 10}
then, running the code above will give:
{5: ['a', 'd'], 1: ['c'], 10: ['b']}

I found that this version is more than 10% faster than the accepted version of a dictionary with 10000 keys.
d = {i: str(i) for i in range(10000)}
new_d = dict(zip(d.values(), d.keys()))

In addition to the other functions suggested above, if you like lambdas:
invert = lambda mydict: {v:k for k, v in mydict.items()}
Or, you could do it this way too:
invert = lambda mydict: dict( zip(mydict.values(), mydict.keys()) )

I think the best way to do this is to define a class. Here is an implementation of a "symmetric dictionary":
class SymDict:
def __init__(self):
self.aToB = {}
self.bToA = {}
def assocAB(self, a, b):
# Stores and returns a tuple (a,b) of overwritten bindings
currB = None
if a in self.aToB: currB = self.bToA[a]
currA = None
if b in self.bToA: currA = self.aToB[b]
self.aToB[a] = b
self.bToA[b] = a
return (currA, currB)
def lookupA(self, a):
if a in self.aToB:
return self.aToB[a]
return None
def lookupB(self, b):
if b in self.bToA:
return self.bToA[b]
return None
Deletion and iteration methods are easy enough to implement if they're needed.
This implementation is way more efficient than inverting an entire dictionary (which seems to be the most popular solution on this page). Not to mention, you can add or remove values from your SymDict as much as you want, and your inverse-dictionary will always stay valid -- this isn't true if you simply reverse the entire dictionary once.

If the values aren't unique, and you're a little hardcore:
inv_map = dict(
(v, [k for (k, xx) in filter(lambda (key, value): value == v, my_map.items())])
for v in set(my_map.values())
)
Especially for a large dict, note that this solution is far less efficient than the answer Python reverse / invert a mapping because it loops over items() multiple times.

This handles non-unique values and retains much of the look of the unique case.
inv_map = {v:[k for k in my_map if my_map[k] == v] for v in my_map.itervalues()}
For Python 3.x, replace itervalues with values.

I am aware that this question already has many good answers, but I wanted to share this very neat solution that also takes care of duplicate values:
def dict_reverser(d):
seen = set()
return {v: k for k, v in d.items() if v not in seen or seen.add(v)}
This relies on the fact that set.add always returns None in Python.

Here is another way to do it.
my_map = {'a': 1, 'b': 2}
inv_map= {}
for key in my_map.keys() :
val = my_map[key]
inv_map[val] = key

dict([(value, key) for key, value in d.items()])

Function is symmetric for values of type list; Tuples are coverted to lists when performing reverse_dict(reverse_dict(dictionary))
def reverse_dict(dictionary):
reverse_dict = {}
for key, value in dictionary.iteritems():
if not isinstance(value, (list, tuple)):
value = [value]
for val in value:
reverse_dict[val] = reverse_dict.get(val, [])
reverse_dict[val].append(key)
for key, value in reverse_dict.iteritems():
if len(value) == 1:
reverse_dict[key] = value[0]
return reverse_dict

Since dictionaries require one unique key within the dictionary unlike values, we have to append the reversed values into a list of sort to be included within the new specific keys.
def r_maping(dictionary):
List_z=[]
Map= {}
for z, x in dictionary.iteritems(): #iterate through the keys and values
Map.setdefault(x,List_z).append(z) #Setdefault is the same as dict[key]=default."The method returns the key value available in the dictionary and if given key is not available then it will return provided default value. Afterward, we will append into the default list our new values for the specific key.
return Map

Fast functional solution for non-bijective maps (values not unique):
from itertools import imap, groupby
def fst(s):
return s[0]
def snd(s):
return s[1]
def inverseDict(d):
"""
input d: a -> b
output : b -> set(a)
"""
return {
v : set(imap(fst, kv_iter))
for (v, kv_iter) in groupby(
sorted(d.iteritems(),
key=snd),
key=snd
)
}
In theory this should be faster than adding to the set (or appending to the list) one by one like in the imperative solution.
Unfortunately the values have to be sortable, the sorting is required by groupby.

Try this for python 2.7/3.x
inv_map={};
for i in my_map:
inv_map[my_map[i]]=i
print inv_map

def invertDictionary(d):
myDict = {}
for i in d:
value = d.get(i)
myDict.setdefault(value,[]).append(i)
return myDict
print invertDictionary({'a':1, 'b':2, 'c':3 , 'd' : 1})
This will provide output as : {1: ['a', 'd'], 2: ['b'], 3: ['c']}

A lambda solution for current python 3.x versions:
d1 = dict(alice='apples', bob='bananas')
d2 = dict(map(lambda key: (d1[key], key), d1.keys()))
print(d2)
Result:
{'apples': 'alice', 'bananas': 'bob'}
This solution does not check for duplicates.
Some remarks:
The lambda construct can access d1 from the outer scope, so we only
pass in the current key. It returns a tuple.
The dict() constructor accepts a list of tuples. It
also accepts the result of a map, so we can skip the conversion to a
list.
This solution has no explicit for loop. It also avoids using a list comprehension for those who are bad at math ;-)

Taking up the highly voted answer starting If the values in my_map aren't unique:, I had a problem where not only the values were not unique, but in addition, they were a list, with each item in the list consisting again of a list of three elements: a string value, a number, and another number.
Example:
mymap['key1'] gives you:
[('xyz', 1, 2),
('abc', 5, 4)]
I wanted to switch only the string value with the key, keeping the two number elements at the same place. You simply need another nested for loop then:
inv_map = {}
for k, v in my_map.items():
for x in v:
# with x[1:3] same as x[1], x[2]:
inv_map[x[0]] = inv_map.get(x[0], []) + [k, x[1:3]]
Example:
inv_map['abc'] now gives you:
[('key1', 1, 2),
('key1', 5, 4)]

This works even if you have non-unique values in the original dictionary.
def dict_invert(d):
'''
d: dict
Returns an inverted dictionary
'''
# Your code here
inv_d = {}
for k, v in d.items():
if v not in inv_d.keys():
inv_d[v] = [k]
else:
inv_d[v].append(k)
inv_d[v].sort()
print(f"{inv_d[v]} are the values")
return inv_d

I would do it that way in python 2.
inv_map = {my_map[x] : x for x in my_map}

Not something completely different, just a bit rewritten recipe from Cookbook. It's futhermore optimized by retaining setdefault method, instead of each time getting it through the instance:
def inverse(mapping):
'''
A function to inverse mapping, collecting keys with simillar values
in list. Careful to retain original type and to be fast.
>> d = dict(a=1, b=2, c=1, d=3, e=2, f=1, g=5, h=2)
>> inverse(d)
{1: ['f', 'c', 'a'], 2: ['h', 'b', 'e'], 3: ['d'], 5: ['g']}
'''
res = {}
setdef = res.setdefault
for key, value in mapping.items():
setdef(value, []).append(key)
return res if mapping.__class__==dict else mapping.__class__(res)
Designed to be run under CPython 3.x, for 2.x replace mapping.items() with mapping.iteritems()
On my machine runs a bit faster, than other examples here

Merge several Python dictionaries [duplicate]

This question already has answers here:
How to merge dicts, collecting values from matching keys?
(17 answers)
Closed 3 months ago.
I have to merge list of python dictionary. For eg:
dicts[0] = {'a':1, 'b':2, 'c':3}
dicts[1] = {'a':1, 'd':2, 'c':'foo'}
dicts[2] = {'e':57,'c':3}
super_dict = {'a':[1], 'b':[2], 'c':[3,'foo'], 'd':[2], 'e':[57]}
I wrote the following code:
super_dict = {}
for d in dicts:
for k, v in d.items():
if super_dict.get(k) is None:
super_dict[k] = []
if v not in super_dict.get(k):
super_dict[k].append(v)
Can it be presented more elegantly / optimized?
Note
I found another question on SO but its about merging exactly 2 dictionaries.

You can iterate over the dictionaries directly -- no need to use range. The setdefault method of dict looks up a key, and returns the value if found. If not found, it returns a default, and also assigns that default to the key.
super_dict = {}
for d in dicts:
for k, v in d.iteritems(): # d.items() in Python 3+
super_dict.setdefault(k, []).append(v)
Also, you might consider using a defaultdict. This just automates setdefault by calling a function to return a default value when a key isn't found.
import collections
super_dict = collections.defaultdict(list)
for d in dicts:
for k, v in d.iteritems(): # d.items() in Python 3+
super_dict[k].append(v)
Also, as Sven Marnach astutely observed, you seem to want no duplication of values in your lists. In that case, set gets you what you want:
import collections
super_dict = collections.defaultdict(set)
for d in dicts:
for k, v in d.iteritems(): # d.items() in Python 3+
super_dict[k].add(v)

from collections import defaultdict
dicts = [{'a':1, 'b':2, 'c':3},
{'a':1, 'd':2, 'c':'foo'},
{'e':57, 'c':3} ]
super_dict = defaultdict(set) # uses set to avoid duplicates
for d in dicts:
for k, v in d.items(): # use d.iteritems() in python 2
super_dict[k].add(v)

you can use this behaviour of dict. (a bit elegant)
a = {'a':1, 'b':2, 'c':3}
b = {'d':1, 'e':2, 'f':3}
c = {1:1, 2:2, 3:3}
merge = {**a, **b, **c}
print(merge) # {'a': 1, 'b': 2, 'c': 3, 'd': 1, 'e': 2, 'f': 3, 1: 1, 2: 2, 3: 3}
and you are good to go :)

Merge the keys of all dicts, and for each key assemble the list of values:
super_dict = {}
for k in set(k for d in dicts for k in d):
super_dict[k] = [d[k] for d in dicts if k in d]
The expression set(k for d in dicts for k in d) builds a set of all unique keys of all dictionaries. For each of these unique keys, we use the list comprehension [d[k] for d in dicts if k in d] to build the list of values from all dicts for this key.
Since you only seem to one the unique value of each key, you might want to use sets instead:
super_dict = {}
for k in set(k for d in dicts for k in d):
super_dict[k] = set(d[k] for d in dicts if k in d)

It seems like most of the answers using comprehensions are not all that readable. In case any gets lost in the mess of answers above this might be helpful (although extremely late...). Just loop over the items of each dict and place them in a separate one.
super_dict = {key:val for d in dicts for key,val in d.items()}

When the value of the keys are in list:
from collections import defaultdict
dicts = [{'a':[1], 'b':[2], 'c':[3]},
{'a':[11], 'd':[2], 'c':['foo']},
{'e':[57], 'c':[3], "a": [1]} ]
super_dict = defaultdict(list) # uses set to avoid duplicates
for d in dicts:
for k, v in d.items(): # use d.iteritems() in python 2
super_dict[k] = list(set(super_dict[k] + v))
combined_dict = {}
for elem in super_dict.keys():
combined_dict[elem] = super_dict[elem]
combined_dict
## output: {'a': [1, 11], 'b': [2], 'c': [3, 'foo'], 'd': [2], 'e': [57]}

I have a very easy to go solution without any imports.
I use the dict.update() method.
But sadly it will overwrite, if same key appears in more than one dictionary, then the most recently merged dict's value will appear in the output.
dict1 = {'Name': 'Zara', 'Age': 7}
dict2 = {'Sex': 'female' }
dict3 = {'Status': 'single', 'Age': 27}
dict4 = {'Occupation':'nurse', 'Wage': 3000}
def mergedict(*args):
output = {}
for arg in args:
output.update(arg)
return output
print(mergedict(dict1, dict2, dict3, dict4))
The output is this:
{'Name': 'Zara', 'Age': 27, 'Sex': 'female', 'Status': 'single', 'Occupation': 'nurse', 'Wage': 3000}

Perhaps a more modern and concise approach for those who use python 3.3 or later versions is the use of ChainMap from the collections module.
from collections import ChainMap
d1 = {'a': 1, 'b': 3}
d2 = {'c': 2}
d3 = {'d': 7, 'a': 9}
d4 = {}
combo = dict(ChainMap(d1, d2, d3, d4))
# {'d': 7, 'a': 1, 'c': 2, 'b': 3}
For a larger collection of dict objects then star operator works
dict(ChainMap(*dict_collection))
Note that the resulting dictionary seems to only keep the value of the first key it encounters in the ordered collection and ignores any further duplicates.

This may be a bit more elegant:
super_dict = {}
for d in dicts:
for k, v in d.iteritems():
l=super_dict.setdefault(k,[])
if v not in l:
l.append(v)
UPDATE: made change suggested by Sven
UPDATE: changed to avoid duplicates (thanks Marcin and Steven)

Never forget that the standard libraries have a wealth of tools for dealing with dicts and iteration:
from itertools import chain
from collections import defaultdict
super_dict = defaultdict(list)
for k,v in chain.from_iterable(d.iteritems() for d in dicts):
if v not in super_dict[k]: super_dict[k].append(v)
Note that the if v not in super_dict[k] can be avoided by using defaultdict(set) as per Steven Rumbalski's answer.

If you assume that the keys in which you are interested are at the same nested level, you can recursively traverse each dictionary and create a new dictionary using that key, effectively merging them.
merged = {}
for d in dicts:
def walk(d,merge):
for key, item in d.items():
if isinstance(item, dict):
merge.setdefault(key, {})
walk(item, merge[key])
else:
merge.setdefault(key, [])
merge[key].append(item)
walk(d,merged)
For example, say you have the following dictionaries you want to merge.
dicts = [{'A': {'A1': {'FOO': [1,2,3]}}},
{'A': {'A1': {'A2': {'BOO': [4,5,6]}}}},
{'A': {'A1': {'FOO': [7,8]}}},
{'B': {'B1': {'COO': [9]}}},
{'B': {'B2': {'DOO': [10,11,12]}}},
{'C': {'C1': {'C2': {'POO':[13,14,15]}}}},
{'C': {'C1': {'ROO': [16,17]}}}]
Using the key at each level, you should get something like this:
{'A': {'A1': {'FOO': [[1, 2, 3], [7, 8]],
'A2': {'BOO': [[4, 5, 6]]}}},
'B': {'B1': {'COO': [[9]]},
'B2': {'DOO': [[10, 11, 12]]}},
'C': {'C1': {'C2': {'POO': [[13, 14, 15]]},
'ROO': [[16, 17]]}}}
Note: I assume the leaf at each branch is a list of some kind, but you can obviously change the logic to do whatever is necessary for your situation.

This is a more recent enhancement over the prior answer by ElbowPipe, using newer syntax introduced in Python 3.9 for merging dictionaries. Note that this answer does not merge conflicting values into a list!
> import functools
> import operator
> functools.reduce(operator.or_, [{0:1}, {2:3, 4:5}, {2:6}])
{0: 1, 2: 6, 4: 5}

For a oneliner, the following could be used:
{key: {d[key] for d in dicts if key in d} for key in {key for d in dicts for key in d}}
although readibility would benefit from naming the combined key set:
combined_key_set = {key for d in dicts for key in d}
super_dict = {key: {d[key] for d in dicts if key in d} for key in combined_key_set}
Elegance can be debated but personally I prefer comprehensions over for loops. :)
(The dictionary and set comprehensions are available in Python 2.7/3.1 and newer.)

python 3.x (reduce is builtin for python 2.x, so no need to import if in 2.x)
import operator
from functools import operator.add
a = [{'a': 1}, {'b': 2}, {'c': 3, 'd': 4}]
dict(reduce(operator.add, map(list,(map(dict.items, a))))
map(dict.items, a) # converts to list of key, value iterators
map(list, ... # converts to iterator equivalent of [[[a, 1]], [[b, 2]], [[c, 3],[d,4]]]
reduce(operator.add, ... # reduces the multiple list down to a single list

My solution is similar to #senderle proposed, but instead of for loop I used map
super_dict = defaultdict(set)
map(lambda y: map(lambda x: super_dict[x].add(y[x]), y), dicts)

The use of defaultdict is good, this also can be done with the use of itertools.groupby.
import itertools
# output all dict items, and sort them by key
dicts_ele = sorted( ( item for d in dicts for item in d.items() ), key = lambda x: x[0] )
# groups items by key
ele_groups = itertools.groupby( dicts_ele, key = lambda x: x[0] )
# iterates over groups and get item value
merged = { k: set( v[1] for v in grouped ) for k, grouped in ele_groups }
and obviously, you can merge this block of code into one-line style
merged = {
k: set( v[1] for v in grouped )
for k, grouped in (
itertools.groupby(
sorted(
( item for d in dicts for item in d.items() ),
key = lambda x: x[0]
),
key = lambda x: x[0]
)
)
}

I'm a bit late to the game but I did it in 2 lines with no dependencies beyond python itself:
flatten = lambda *c: (b for a in c for b in (flatten(*a) if isinstance(a, (tuple, list)) else (a,)))
o = reduce(lambda d1,d2: dict((k, list(flatten([d1.get(k), d2.get(k)]))) for k in set(d1.keys() + d2.keys())), dicts)
# output:
# {'a': [1, 1, None], 'c': [3, 'foo', 3], 'b': [2, None, None], 'e': [None, 57], 'd': [None, 2, None]}
Though if you don't care about nested lists, then:
o2 = reduce(lambda d1,d2: dict((k, [d1.get(k), d2.get(k)]) for k in set(d1.keys() + d2.keys())), dicts)
# output:
# {'a': [[1, 1], None], 'c': [[3, 'foo'], 3], 'b': [[2, None], None], 'e': [None, 57], 'd': [[None, 2], None]}

Reverse / invert a dictionary mapping

Given a dictionary like so:
my_map = {'a': 1, 'b': 2}
How can one invert this map to get:
inv_map = {1: 'a', 2: 'b'}

Python 3+:
inv_map = {v: k for k, v in my_map.items()}
Python 2:
inv_map = {v: k for k, v in my_map.iteritems()}

Assuming that the values in the dict are unique:
Python 3:
dict((v, k) for k, v in my_map.items())
Python 2:
dict((v, k) for k, v in my_map.iteritems())

If the values in my_map aren't unique:
Python 3:
inv_map = {}
for k, v in my_map.items():
inv_map[v] = inv_map.get(v, []) + [k]
Python 2:
inv_map = {}
for k, v in my_map.iteritems():
inv_map[v] = inv_map.get(v, []) + [k]

To do this while preserving the type of your mapping (assuming that it is a dict or a dict subclass):
def inverse_mapping(f):
return f.__class__(map(reversed, f.items()))

Try this:
inv_map = dict(zip(my_map.values(), my_map.keys()))
(Note that the Python docs on dictionary views explicitly guarantee that .keys() and .values() have their elements in the same order, which allows the approach above to work.)
Alternatively:
inv_map = dict((my_map[k], k) for k in my_map)
or using python 3.0's dict comprehensions
inv_map = {my_map[k] : k for k in my_map}

Another, more functional, way:
my_map = { 'a': 1, 'b':2 }
dict(map(reversed, my_map.items()))

We can also reverse a dictionary with duplicate keys using defaultdict:
from collections import Counter, defaultdict
def invert_dict(d):
d_inv = defaultdict(list)
for k, v in d.items():
d_inv[v].append(k)
return d_inv
text = 'aaa bbb ccc ddd aaa bbb ccc aaa'
c = Counter(text.split()) # Counter({'aaa': 3, 'bbb': 2, 'ccc': 2, 'ddd': 1})
dict(invert_dict(c)) # {1: ['ddd'], 2: ['bbb', 'ccc'], 3: ['aaa']}
See here:
This technique is simpler and faster than an equivalent technique using dict.setdefault().

This expands upon the answer by Robert, applying to when the values in the dict aren't unique.
class ReversibleDict(dict):
# Ref: https://stackoverflow.com/a/13057382/
def reversed(self):
"""
Return a reversed dict, with common values in the original dict
grouped into a list in the returned dict.
Example:
>>> d = ReversibleDict({'a': 3, 'c': 2, 'b': 2, 'e': 3, 'd': 1, 'f': 2})
>>> d.reversed()
{1: ['d'], 2: ['c', 'b', 'f'], 3: ['a', 'e']}
"""
revdict = {}
for k, v in self.items():
revdict.setdefault(v, []).append(k)
return revdict
The implementation is limited in that you cannot use reversed twice and get the original back. It is not symmetric as such. It is tested with Python 2.6. Here is a use case of how I am using to print the resultant dict.
If you'd rather use a set than a list, and there could exist unordered applications for which this makes sense, instead of setdefault(v, []).append(k), use setdefault(v, set()).add(k).

Combination of list and dictionary comprehension. Can handle duplicate keys
{v:[i for i in d.keys() if d[i] == v ] for k,v in d.items()}

A case where the dictionary values is a set. Like:
some_dict = {"1":{"a","b","c"},
"2":{"d","e","f"},
"3":{"g","h","i"}}
The inverse would like:
some_dict = {vi: k for k, v in some_dict.items() for vi in v}
The output is like this:
{'c': '1',
'b': '1',
'a': '1',
'f': '2',
'd': '2',
'e': '2',
'g': '3',
'h': '3',
'i': '3'}

For instance, you have the following dictionary:
my_dict = {'a': 'fire', 'b': 'ice', 'c': 'fire', 'd': 'water'}
And you wanna get it in such an inverted form:
inverted_dict = {'fire': ['a', 'c'], 'ice': ['b'], 'water': ['d']}
First Solution. For inverting key-value pairs in your dictionary use a for-loop approach:
# Use this code to invert dictionaries that have non-unique values
inverted_dict = dict()
for key, value in my_dict.items():
inverted_dict.setdefault(value, list()).append(key)
Second Solution. Use a dictionary comprehension approach for inversion:
# Use this code to invert dictionaries that have unique values
inverted_dict = {value: key for key, value in my_dict.items()}
Third Solution. Use reverting the inversion approach (relies on the second solution):
# Use this code to invert dictionaries that have lists of values
my_dict = {value: key for key in inverted_dict for value in my_map[key]}

Lot of answers but didn't find anything clean in case we are talking about a dictionary with non-unique values.
A solution would be:
from collections import defaultdict
inv_map = defaultdict(list)
for k, v in my_map.items():
inv_map[v].append(k)
Example:
If initial dict my_map = {'c': 1, 'd': 5, 'a': 5, 'b': 10}
then, running the code above will give:
{5: ['a', 'd'], 1: ['c'], 10: ['b']}

I found that this version is more than 10% faster than the accepted version of a dictionary with 10000 keys.
d = {i: str(i) for i in range(10000)}
new_d = dict(zip(d.values(), d.keys()))

In addition to the other functions suggested above, if you like lambdas:
invert = lambda mydict: {v:k for k, v in mydict.items()}
Or, you could do it this way too:
invert = lambda mydict: dict( zip(mydict.values(), mydict.keys()) )

I think the best way to do this is to define a class. Here is an implementation of a "symmetric dictionary":
class SymDict:
def __init__(self):
self.aToB = {}
self.bToA = {}
def assocAB(self, a, b):
# Stores and returns a tuple (a,b) of overwritten bindings
currB = None
if a in self.aToB: currB = self.bToA[a]
currA = None
if b in self.bToA: currA = self.aToB[b]
self.aToB[a] = b
self.bToA[b] = a
return (currA, currB)
def lookupA(self, a):
if a in self.aToB:
return self.aToB[a]
return None
def lookupB(self, b):
if b in self.bToA:
return self.bToA[b]
return None
Deletion and iteration methods are easy enough to implement if they're needed.
This implementation is way more efficient than inverting an entire dictionary (which seems to be the most popular solution on this page). Not to mention, you can add or remove values from your SymDict as much as you want, and your inverse-dictionary will always stay valid -- this isn't true if you simply reverse the entire dictionary once.

If the values aren't unique, and you're a little hardcore:
inv_map = dict(
(v, [k for (k, xx) in filter(lambda (key, value): value == v, my_map.items())])
for v in set(my_map.values())
)
Especially for a large dict, note that this solution is far less efficient than the answer Python reverse / invert a mapping because it loops over items() multiple times.

This handles non-unique values and retains much of the look of the unique case.
inv_map = {v:[k for k in my_map if my_map[k] == v] for v in my_map.itervalues()}
For Python 3.x, replace itervalues with values.

I am aware that this question already has many good answers, but I wanted to share this very neat solution that also takes care of duplicate values:
def dict_reverser(d):
seen = set()
return {v: k for k, v in d.items() if v not in seen or seen.add(v)}
This relies on the fact that set.add always returns None in Python.

Here is another way to do it.
my_map = {'a': 1, 'b': 2}
inv_map= {}
for key in my_map.keys() :
val = my_map[key]
inv_map[val] = key

dict([(value, key) for key, value in d.items()])

Function is symmetric for values of type list; Tuples are coverted to lists when performing reverse_dict(reverse_dict(dictionary))
def reverse_dict(dictionary):
reverse_dict = {}
for key, value in dictionary.iteritems():
if not isinstance(value, (list, tuple)):
value = [value]
for val in value:
reverse_dict[val] = reverse_dict.get(val, [])
reverse_dict[val].append(key)
for key, value in reverse_dict.iteritems():
if len(value) == 1:
reverse_dict[key] = value[0]
return reverse_dict

Since dictionaries require one unique key within the dictionary unlike values, we have to append the reversed values into a list of sort to be included within the new specific keys.
def r_maping(dictionary):
List_z=[]
Map= {}
for z, x in dictionary.iteritems(): #iterate through the keys and values
Map.setdefault(x,List_z).append(z) #Setdefault is the same as dict[key]=default."The method returns the key value available in the dictionary and if given key is not available then it will return provided default value. Afterward, we will append into the default list our new values for the specific key.
return Map

Fast functional solution for non-bijective maps (values not unique):
from itertools import imap, groupby
def fst(s):
return s[0]
def snd(s):
return s[1]
def inverseDict(d):
"""
input d: a -> b
output : b -> set(a)
"""
return {
v : set(imap(fst, kv_iter))
for (v, kv_iter) in groupby(
sorted(d.iteritems(),
key=snd),
key=snd
)
}
In theory this should be faster than adding to the set (or appending to the list) one by one like in the imperative solution.
Unfortunately the values have to be sortable, the sorting is required by groupby.

Try this for python 2.7/3.x
inv_map={};
for i in my_map:
inv_map[my_map[i]]=i
print inv_map

def invertDictionary(d):
myDict = {}
for i in d:
value = d.get(i)
myDict.setdefault(value,[]).append(i)
return myDict
print invertDictionary({'a':1, 'b':2, 'c':3 , 'd' : 1})
This will provide output as : {1: ['a', 'd'], 2: ['b'], 3: ['c']}

A lambda solution for current python 3.x versions:
d1 = dict(alice='apples', bob='bananas')
d2 = dict(map(lambda key: (d1[key], key), d1.keys()))
print(d2)
Result:
{'apples': 'alice', 'bananas': 'bob'}
This solution does not check for duplicates.
Some remarks:
The lambda construct can access d1 from the outer scope, so we only
pass in the current key. It returns a tuple.
The dict() constructor accepts a list of tuples. It
also accepts the result of a map, so we can skip the conversion to a
list.
This solution has no explicit for loop. It also avoids using a list comprehension for those who are bad at math ;-)

Taking up the highly voted answer starting If the values in my_map aren't unique:, I had a problem where not only the values were not unique, but in addition, they were a list, with each item in the list consisting again of a list of three elements: a string value, a number, and another number.
Example:
mymap['key1'] gives you:
[('xyz', 1, 2),
('abc', 5, 4)]
I wanted to switch only the string value with the key, keeping the two number elements at the same place. You simply need another nested for loop then:
inv_map = {}
for k, v in my_map.items():
for x in v:
# with x[1:3] same as x[1], x[2]:
inv_map[x[0]] = inv_map.get(x[0], []) + [k, x[1:3]]
Example:
inv_map['abc'] now gives you:
[('key1', 1, 2),
('key1', 5, 4)]

This works even if you have non-unique values in the original dictionary.
def dict_invert(d):
'''
d: dict
Returns an inverted dictionary
'''
# Your code here
inv_d = {}
for k, v in d.items():
if v not in inv_d.keys():
inv_d[v] = [k]
else:
inv_d[v].append(k)
inv_d[v].sort()
print(f"{inv_d[v]} are the values")
return inv_d

I would do it that way in python 2.
inv_map = {my_map[x] : x for x in my_map}

Not something completely different, just a bit rewritten recipe from Cookbook. It's futhermore optimized by retaining setdefault method, instead of each time getting it through the instance:
def inverse(mapping):
'''
A function to inverse mapping, collecting keys with simillar values
in list. Careful to retain original type and to be fast.
>> d = dict(a=1, b=2, c=1, d=3, e=2, f=1, g=5, h=2)
>> inverse(d)
{1: ['f', 'c', 'a'], 2: ['h', 'b', 'e'], 3: ['d'], 5: ['g']}
'''
res = {}
setdef = res.setdefault
for key, value in mapping.items():
setdef(value, []).append(key)
return res if mapping.__class__==dict else mapping.__class__(res)
Designed to be run under CPython 3.x, for 2.x replace mapping.items() with mapping.iteritems()
On my machine runs a bit faster, than other examples here