Reputation: 103
Python: Replace values in nested dictionary
I want to replace the values (formated as strings) with the same values as integers, whenever the key is 'current_values'.
d = {'id': '10', 'datastreams': [{'current_value': '5'}, {'current_value': '4'}]}
Desired Output:
d = {'id': '10', 'datastreams': [{'current_value': 5}, {'current_value': 4}]}
Upvotes: 6
Views: 13278
Answers (8)
Reputation: 3016
Based on Cloudkollektiv 's excellent idea, I've refactored the code which basically can do the same as the original functions, but all in one.
I've also added the possibility, instead of replacing an orignal string by a replacement string, to replace every value by fn(value) where fn is a function given as argument to original.
[EDIT] I've improved the code so it retains the initial object type, which comes in handy when dealing with structs like CommentedMap from ruamel.yaml
I've also added an option to replace all values by fn(key, value) for dicts where fn is a function given as argument to original.
[/EDIT]
def replace_in_iterable(
src: Union[dict, list],
original: Union[str, Callable],
replacement: Any = None,
callable_wants_key: bool = False,
):
"""
Recursive replace data in a struct
Replaces every instance of string original with string replacement in a list/dict
If original is a callable function, it will replace every value with original(value)
If original is a callable function and callable_wants_key == True,
it will replace every value with original(key, value) for dicts
and with original(value) for any other data types
"""
def _replace_in_iterable(key, _src):
if isinstance(_src, dict) or isinstance(_src, list):
_src = replace_in_iterable(_src, original, replacement, callable_wants_key)
elif isinstance(original, Callable):
if callable_wants_key:
_src = original(key, _src)
else:
_src = original(_src)
elif isinstance(_src, str) and isinstance(replacement, str):
_src = _src.replace(original, replacement)
else:
_src = replacement
return _src
if isinstance(src, dict):
for key, value in src.items():
src[key] = _replace_in_iterable(key, value)
elif isinstance(src, list):
result = []
for entry in src:
result.append(_replace_in_iterable(None, entry))
src = result
else:
src = _replace_in_iterable(None, src)
return src
TL;DR: You can directly use this function as package via pip with:
pip install ofunctions.misc
Then use it with
from ofunctions.misc import replace_in_iterable
def test(string):
return f"-{string}-"
output = replace_in_iterable(input, test)
Input
input = {
'key1': 'a string',
'key2': 'another string',
'key3': [
'a string',
'another string',
[1, 2, 3],
{
'key1': 'a string',
'key2': 'another string'
}
],
'key4': {
'key1': 'a string',
'key2': 'another string',
'key3': [
'a string',
'another string',
500,
1000
]
},
'key5': {
'key1': [
{
'key1': 'a string'
}
]
}
}
Output
input = {
'key1': '-a string-',
'key2': '-another string-',
'key3': [
'-a string-',
'-another string-',
['-1-', '-2-', '-3-'],
{
'key1': '-a string-',
'key2': '-another string-'
}
],
'key4': {
'key1': '-a string-',
'key2': '-another string-',
'key3': [
'-a string-',
'-another string-',
'-500-',
'-1000-'
]
},
'key5': {
'key1': [
{
'key1': '-a string-'
}
]
}
}
Of course the original syntax via output = replace_in_iterable(input, "string", "something) still works.
Upvotes: 0
Reputation: 14699
The following piece of code replaces (substrings of) values in a dictionary. It works for nested json structures and copes with json, list and string types. You can easily add other types if needed.
def dict_replace_value(d: dict, old: str, new: str) -> dict:
x = {}
for k, v in d.items():
if isinstance(v, dict):
v = dict_replace_value(v, old, new)
elif isinstance(v, list):
v = list_replace_value(v, old, new)
elif isinstance(v, str):
v = v.replace(old, new)
x[k] = v
return x
def list_replace_value(l: list, old: str, new: str) -> list:
x = []
for e in l:
if isinstance(e, list):
e = list_replace_value(e, old, new)
elif isinstance(e, dict):
e = dict_replace_value(e, old, new)
elif isinstance(e, str):
e = e.replace(old, new)
x.append(e)
return x
# See input and output below
output = dict_replace_value(input, 'string', 'something')
Input:
input = {
'key1': 'a string',
'key2': 'another string',
'key3': [
'a string',
'another string',
[1, 2, 3],
{
'key1': 'a string',
'key2': 'another string'
}
],
'key4': {
'key1': 'a string',
'key2': 'another string',
'key3': [
'a string',
'another string',
500,
1000
]
},
'key5': {
'key1': [
{
'key1': 'a string'
}
]
}
}
Output:
print(output)
{
"key1":"a something",
"key2":"another something",
"key3":[
"a something",
"another something",
[
1,
2,
3
],
{
"key1":"a something",
"key2":"another something"
}
],
"key4":{
"key1":"a something",
"key2":"another something",
"key3":[
"a something",
"another something",
500,
1000
]
},
"key5":{
"key1":[
{
"key1":"a something"
}
]
}
}
Upvotes: 26
Reputation: 1
Taking alec_djinn's solution little farther to handle also nested dicts:
def f(d):
for k,v in d.items():
if k == 'current_value':
d[k] = int(v)
elif type(v) is list:
for item in v:
if type(item) is dict:
f(item)
if type(v) is dict:
f(v)
Upvotes: 0
Reputation: 16772
d = {'id': '10', 'datastreams': [{'current_value': '5'}, {'current_value': '4'}]}
for elem in d['datastreams']: # for each elem in the list datastreams
for k,v in elem.items(): # for key,val in the elem of the list
if 'current_value' in k: # if current_value is in the key
elem[k] = int(v) # Cast it to int
print(d)
OUTPUT:
{'id': '10', 'datastreams': [{'current_value': 5}, {'current_value': 4}]}
Upvotes: 5
Reputation: 10789
A general approach (assuming you don't know in advance which key of the dict is pointing to a list) would be to iterate over the dict and check the type of its values and then iterate again into each value if needed.
In your case, your dictionary may contain a list of dictionaries as values, so it is enough to check if a value is of type list, if so, iterate over the list and change the dicts you need.
It can be done recursively with a function like the following:
def f(d):
for k,v in d.items():
if k == 'current_value':
d[k] = int(v)
elif type(v) is list:
for item in v:
if type(item) is dict:
f(item)
>>> d = {'id': '10', 'datastreams': [{'current_value': '5'}, {'current_value': '4'}]}
>>> f(d)
>>> d
{'id': '10', 'datastreams': [{'current_value': 5}, {'current_value': 4}]}
Upvotes: 2
Reputation: 41
You could use this method which would loop through checks for current_value in list and change it to integer by passing the value through int() function:
for value in d.values():
for element in value:
if 'current_value' in element:
element['current_value'] = int(element['current_value'])
Upvotes: 0
Reputation: 5463
You can use ast.literal_eval to evaluate the underlying value for items with current_value key in the d['datastreams'] list. Then check whether the type is an int using isinstance for such values. Finally, type cast such values to int.
import ast
d = {'id': '10', 'datastreams': [{'current_value': '5'}, {'current_value': '4'}]}
for i in d['datastreams']:
for k,v in i.items():
if 'current_value' in k and isinstance(ast.literal_eval(v),int):
i[k] = int(v)
#Output:
print(d)
{'id': '10', 'datastreams': [{'current_value': 5}, {'current_value': 4}]}
Upvotes: 0
Reputation: 5833
Can be done with list comprehension:
d['datastreams'] = [{'current_value': int(ds['current_value'])} if ('current_value' in ds) else ds for ds in d['datastreams']]
Upvotes: 1
Related Questions
- Python dictionary replace nested values
- Replace keys in a nested dictionary
- How to replace values in nested dictionaries by new values in a list in python
- Loop through nested dictionary, replacing values with items from a list
- How can I replace a key-value pair in a nested dictionary with the value from the same key-value pair?
- How to replace values in a nested dict
- how to use replace for replacing keys in a nested dictionary
- Replace elements in a nested dict with the appropriate in a list
- python: replace strings in a nested dictionary
- Replacing values in a nested dictionary