Develop Reference

How to make a dictionary from another dictionary values

I have a list of dictionaries, something like this
input = [
{"0_question": "how are you"},
{"0_answer": "good"},
{"4_question": "what time is it"},
{"4_answer": "morning"},
]
and I want this:
expected_result = {
0: {"how are you": "good"},
4: {"what time is it" : "morning"},
}
this is what I tried:
x = {}
l= []
for i, d in enumerate(a):
for k, v in d.items():
l.append(v)
x[l[i]] = l[i+1]

Split the original dict key into numeric and type part:
a = [
{
"0_question": "how are you"
}, {
"0_answer": "good"
}, {
"4_question": "what time is it"
}, {
"4_answer": "morning"
}
]
dic = {}
for item in a:
value = list(item.values())[0]
num, itemType = list(item.keys())[0].split('_')
num = int(num)
if itemType == 'question':
dic[num] = {value: ''}
else:
key = list(dic[num].keys())[0]
dic[num][key] = value
print(dic)
Out:
{0: {'how are you': 'good'}, 4: {'what time is it': 'morning'}}

You can use a nested loop and the modulo operator (%). Use regex to find the question/answer number:
import re
l = [
{"0_question": "how are you"},
{"0_answer": "good"},
{"4_question": "what time is it"},
{"4_answer": "morning"},
{"157_question": "which fruit do you want"},
{"157_answer": "apple"}
]
new_d = {}
for i, d in enumerate(l):
for k, v in d.items():
result = re.search("\d+", k).group(0)
num = int(result)
if ((i + 1) % 2) == 0:
for new_k in new_d[num].keys():
new_d[num][new_k] = v
else:
new_d[num] = {v: ""}
print(new_d)
Output
{0: {'how are you': 'good'}, 4: {'what time is it': 'morning'}, 157: {'which fruit do you want': 'apple'}}

Use This Code
a = [{"0_question": "how are you"},
{"0_answer": "good"},
{"4_question": "what time is it"},
{"4_answer": "morning"}]
dic = {}
for i in range(len(a) - 1):
if i % 2 == 0:
dic[list(a[i].keys())[0]] = {list(a[i].values())[0]: list(a[i + 1].values())[0]}
print(dic)

This has fewer lines and hopefully increases readability. The downside is it uses more resources by restructuring the initial data.
sample_data = [{"0_question" : "how are you"},
{"0_answer": "good"},
{"4_question" : "what time is it"},
{"4_answer": "morning"}]
restructured_data = {k: v for data in sample_data for k, v in data.items()}
nums = list(set([int(k.split("_")[0]) for k, v in restructured_data.items()]))
new_dict = {num: {restructured_data[f'{num}_question']: restructured_data[f'{num}_answer']} for num in nums}
print(new_dict)
output:
{0: {'how are you': 'good'}, 4: {'what time is it': 'morning'}}

data = [{"0_question" : "how are you"},
{"0_answer": "good"},
{"4_question" : "what time is it"},
{"4_answer": "morning"}]
# unpack dict values into any iterable data structure
# list in my case
unpacked_data = list(map(lambda dict_: list(dict_.items()), data))
# dict_items is not subscriptable, so unpack it again
# also exctracting number from key.
processed_data = list(
map(
lambda values_tuple: (values_tuple[0][0][0], values_tuple[0][1]),
unpacked_data
)
)
result_dict = {}
# according to input data we assuame that there is a pairs of (question: answer).
# using zip and slicing for pair iteration
for question, answer in zip(processed_data[0::2], processed_data[1::2]):
result_dict[question[0]] = {question[1] : answer[1]}
print(result_dict)
UPD: this may be not the most optimal solution, but i hope that is pretty clear
for understanding

Check Below Code:
{ int(list(i1)[0][0]) : {list(i1.values())[0] : list(i2.values())[0] } for i1, i2 in zip(input[::2], input[1::2])}
Output:

Create nested dictionary from keys seperated by dot(.) in python

I have a requirement where I have a keys in string format combined by dot(.) and the value associated with that string of key and I want to create a dictionary.
key1 = "A.B.C.D"
text_to_be_inserted1_for_key1 = "Test1"
key2 = "A.B.C.E"
text_to_be_inserted_for_key2 = "Test2"
Expected result
dict = {
"A": {
"B" : {
"C" : {
"D" : text_to_be_inserted1_for_key1,
"E" : text_to_be_inserted_for_key2
}
}
}
}

from collections import defaultdict
def deep_dict():
return defaultdict(deep_dict)
result = deep_dict()
def deep_insert(key, value):
d = result
keys = key.split(".")
for subkey in keys[:-1]:
d = d[subkey]
d[keys[-1]] = value
deep_insert("A.B.C.D", "Test1")
deep_insert("A.B.C.E", "Test2")
import json
print(json.dumps(result, indent=4))

You may
for each letter except the last one, create a mapping with the key and a dict
for the last letter create the mapping with the value
def insert(keys, values):
res = {}
for k, v in zip(keys, values):
res_tmp = res
levels = k.split(".")
for level in levels[:-1]:
res_tmp = res_tmp.setdefault(level, {})
res_tmp[levels[-1]] = v
return res
Use
key1 = "A.B.C.D"
value_key1 = "Test1"
key2 = "A.B.C.E"
value_key2 = "Test2"
result = insert([key1, key2], [value_key1, value_key2])
print(result) # {'A': {'B': {'C': {'D': 'Test1', 'E': 'Test2'}}}}

You can solve for each case and then merge
from copy import deepcopy
def dict_of_dicts_merge(x, y):
z = {}
overlapping_keys = x.keys() & y.keys()
for key in overlapping_keys:
z[key] = dict_of_dicts_merge(x[key], y[key])
for key in x.keys() - overlapping_keys:
z[key] = deepcopy(x[key])
for key in y.keys() - overlapping_keys:
z[key] = deepcopy(y[key])
return z
key1 = "A.B.C.D"
text_to_be_inserted_for_key1 = "Test1"
key2 = "A.B.C.E"
text_to_be_inserted_for_key2 = "Test2"
dict1 = {}
newdict = {}
olddict = {}
keys_for_1 = key1.split(".")
keys_for_1.reverse()
olddict[keys_for_1[0]] = text_to_be_inserted_for_key1
for i in range (1,len(keys_for_1)):
newdict = {}
newdict[keys_for_1[i]] = olddict
olddict = newdict
save1 = newdict
newdict = {}
olddict = {}
keys_for_2 = key2.split(".")
keys_for_2.reverse()
olddict[keys_for_2[0]] = text_to_be_inserted_for_key2
for i in range (1,len(keys_for_2)):
newdict = {}
newdict[keys_for_2[i]] = olddict
olddict = newdict
save2 = newdict
dict1 = dict_of_dicts_merge(save1,save2)
print (dict1)

How to deal with columns in pandas dataframe?

I want to do something with column data which is a list. like:
inputs:
col-A
[{'name':'1','age':'12'}, {'name':'2','age':'12'}]
[{'name':'3','age':'18'}, {'name':'7','age':'15'}]
....
outputs:
col-A
[{'1-age':'12'}, {'2-age':'12'}]
[{'3-age':'18'}, {'7-age':'15'}]
....
My code is:
def deal(dict_col, prefix_key):
key_value = dict_col[prefix_key]+'-'
dict_col.pop(prefix_key, None)
items = copy.deepcopy(dict_col)
for key, value in items.items():
dict_col[key_value+key] = dict_col.pop(key)
return dict_col
prefix = "name"
[[deal(sub_item, prefix) for sub_item in item] for item in df[col-A]]
Some items will be processed multiple times.
Because the return value of deal method will be swapped to item in real time?
For example:
For deal method we
input:
{'name':'1','age':'12'}
output:
{'1-age':'12'}
Then the next input may be {'1-age':'12'} , and now we have no name or age to deal with.
How to solve this problem?

You can use the pandas apply method for it here some code:
import pandas as pd
d = {'col-A' : [[{'name' : '1', 'age': '12'}, {'name' : '2', 'age': '12'}],[{'name' : '3', 'age': '18'},{'name' : '7', 'age': '15'}]]}
df = pd.DataFrame(d)
def deal(row, prefix):
out_list = []
for sub_dict in row:
out_dict = {}
out_str = sub_dict.get(prefix) + '-'
for k,v in sub_dict.items():
out_dict[out_str + k] = v
out_list.append(out_dict)
return out_list
prefix = 'name'
df['col-A'] = df['col-A'].apply(lambda x : deal(x, prefix))
print(df)
You could push some of the code in a one-liner if you like that more:
def deal(row, prefix):
out_list = []
for sub_dict in row:
out_dict = dict((sub_dict[prefix] + '-' + k , sub_dict[k]) for k in sub_dict.keys() if k != prefix)
out_list.append(out_dict)
return out_list
prefix = 'name'
df['col-A'] = df['col-A'].apply(lambda x : deal(x, prefix)
Just for the fun of it you could even bring it down to one single line (not recommended due to poor readability:
prefix = "name"
df['col-A'] = df['col-A'].apply(lambda row : [dict((sub_dict[prefix] + '-' + k , sub_dict[k]) for k in sub_dict.keys() if k != prefix) for sub_dict in row])

I believe you need .get function for select with default value if not exist key in dict:
def deal(dict_col, prefix_key):
key_value = dict_col.get(prefix_key, 'not_exist')+'-'
dict_col.pop(prefix_key, None)
items = copy.deepcopy(dict_col)
for key, value in items.items():
dict_col[key_value+key] = dict_col.pop(key)
return dict_col

Python multilevel dict to strings

I have a python dictionary and a dictionary with in some of the values. I'm trying to generate a dotted delimited string of the keys in the structure with the value at the end. With the example below I'd want FIELD0 1 and NAME. I could create a for loop to process the data or a recursive function. I didn't know if there was something prebuilt method for collapsing a multilevel dictionary to delimited strings?
I was trying the following but as you know it will just append the sub dictionaries.
'.'.join('%s %s\n' % i for i in a.items())
{'BOGUS1': 'BOGUS_VAL1',
'BOGUS2': 'BOGUS_VAL1',
'FIELD0': {'F0_VAL1': 1, 'F0_VAL2': 2},
'FIELD1': {'F1_VAL1': 80, 'F1_VAL2': 67, 'F1_VAL3': 100},
'FOOBAR1': 'FB_VAL1',
'NAME': 'VALUE'}
BOGUS2.BOGUS_VAL1
.NAME.VALUE
.BOGUS1.BOGUS_VAL1
.FIELD0.{'F0_VAL1': 1, 'F0_VAL2': 2}
.FIELD1.{'F1_VAL2': 67, 'F1_VAL3': 100, 'F1_VAL1': 80}
.FOOBAR1.FB_VAL1
# Wanted results
FIELD0.F0_VAL1 1
FIELD0.F0_VAL2 2
FIELD1.F1_VAL1 80
FIELD1.F2_VAL1 67
FIELD1.F3_VAL1 100
NAME VALUE

How about something like this:
def dotnotation(d, prefix = ''):
for k, v in d.items():
if type(v) == type(dict()):
dotnotation(v, prefix + str(k) + '.')
else:
print prefix + str(k) + ' = ' + str(v)
Also the formatting can be changed according to the stored types. This should work with your example.

Here is my approach:
def dotted_keys(dic):
""" Generated dot notation keys from a dictionary """
queue = [(None, dic)] # A queue of (prefix, object)
while queue:
prefix, current = queue.pop(0)
for k, v in current.iteritems():
if isinstance(v, dict):
queue.append((k, v))
elif prefix:
yield prefix + '.' + k
else:
yield k
def dict_search(dic, dotted_key, default=None):
""" Take a dictionary and a dotted key and return the value. If not
found, return the value specified by the default parameter.
Example: dict_search(d, 'FIELD0.F0_VAL2')
"""
current = dic
keys = dotted_key.split('.')
for k in keys:
if k in current:
current = current[k]
else:
return default
return current
if __name__ == '__main__':
d = {
'BOGUS1': 'BOGUS_VAL1',
'BOGUS2': 'BOGUS_VAL1',
'FIELD0': {'F0_VAL1': 1, 'F0_VAL2': 2, 'XYZ': {'X1': 9}},
'FIELD1': {'F1_VAL1': 80, 'F1_VAL2': 67, 'F1_VAL3': 100},
'FOOBAR1': 'FB_VAL1',
'NAME': 'VALUE'
}
for k in dotted_keys(d):
print(k, '=', dict_search(d, k))
Output:
BOGUS2 = BOGUS_VAL1
NAME = VALUE
BOGUS1 = BOGUS_VAL1
FOOBAR1 = FB_VAL1
FIELD0.F0_VAL1 = 1
FIELD0.F0_VAL2 = 2
FIELD1.F1_VAL2 = 67
FIELD1.F1_VAL3 = 100
FIELD1.F1_VAL1 = 80
XYZ.X1 = None
The dotted_keys function generates a list of keys in dotted notation while the dict_search function takes a dotted key and return a value.

best way to parse a line in python to a dictionary

I have a file with lines like
account = "TEST1" Qty=100 price = 20.11 subject="some value" values="3=this, 4=that"
There is no special delimiter and each key has a value that is surrounded by double quotes if its a string but not if it is a number. There is no key without a value though there may exist blank strings which are represented as "" and there is no escape character for a quote as it is not needed
I want to know what is a good way to parse this kind of line with python and store the values as key-value pairs in a dictionary

We're going to need a regex for this.
import re, decimal
r= re.compile('([^ =]+) *= *("[^"]*"|[^ ]*)')
d= {}
for k, v in r.findall(line):
if v[:1]=='"':
d[k]= v[1:-1]
else:
d[k]= decimal.Decimal(v)
>>> d
{'account': 'TEST1', 'subject': 'some value', 'values': '3=this, 4=that', 'price': Decimal('20.11'), 'Qty': Decimal('100.0')}
You can use float instead of decimal if you prefer, but it's probably a bad idea if money is involved.

Maybe a bit simpler to follow is the pyparsing rendition:
from pyparsing import *
# define basic elements - use re's for numerics, faster than easier than
# composing from pyparsing objects
integer = Regex(r'[+-]?\d+')
real = Regex(r'[+-]?\d+\.\d*')
ident = Word(alphanums)
value = real | integer | quotedString.setParseAction(removeQuotes)
# define a key-value pair, and a configline as one or more of these
# wrap configline in a Dict so that results are accessible by given keys
kvpair = Group(ident + Suppress('=') + value)
configline = Dict(OneOrMore(kvpair))
src = 'account = "TEST1" Qty=100 price = 20.11 subject="some value" ' \
'values="3=this, 4=that"'
configitems = configline.parseString(src)
Now you can access your pieces using the returned configitems ParseResults object:
>>> print configitems.asList()
[['account', 'TEST1'], ['Qty', '100'], ['price', '20.11'],
['subject', 'some value'], ['values', '3=this, 4=that']]
>>> print configitems.asDict()
{'account': 'TEST1', 'Qty': '100', 'values': '3=this, 4=that',
'price': '20.11', 'subject': 'some value'}
>>> print configitems.dump()
[['account', 'TEST1'], ['Qty', '100'], ['price', '20.11'],
['subject', 'some value'], ['values', '3=this, 4=that']]
- Qty: 100
- account: TEST1
- price: 20.11
- subject: some value
- values: 3=this, 4=that
>>> print configitems.keys()
['account', 'subject', 'values', 'price', 'Qty']
>>> print configitems.subject
some value

A recursive variation of bobince's parses values with embedded equals as dictionaries:
>>> import re
>>> import pprint
>>>
>>> def parse_line(line):
... d = {}
... a = re.compile(r'\s*(\w+)\s*=\s*("[^"]*"|[^ ,]*),?')
... float_re = re.compile(r'^\d.+$')
... int_re = re.compile(r'^\d+$')
... for k,v in a.findall(line):
... if int_re.match(k):
... k = int(k)
... if v[-1] == '"':
... v = v[1:-1]
... if '=' in v:
... d[k] = parse_line(v)
... elif int_re.match(v):
... d[k] = int(v)
... elif float_re.match(v):
... d[k] = float(v)
... else:
... d[k] = v
... return d
...
>>> line = 'account = "TEST1" Qty=100 price = 20.11 subject="some value" values=
"3=this, 4=that"'
>>> pprint.pprint(parse_line(line))
{'Qty': 100,
'account': 'TEST1',
'price': 20.109999999999999,
'subject': 'some value',
'values': {3: 'this', 4: 'that'}}

If you don't want to use a regex, another option is just to read the string a character at a time:
string = 'account = "TEST1" Qty=100 price = 20.11 subject="some value" values="3=this, 4=that"'
inside_quotes = False
key = None
value = ""
dict = {}
for c in string:
if c == '"':
inside_quotes = not inside_quotes
elif c == '=' and not inside_quotes:
key = value
value = ''
elif c == ' ':
if inside_quotes:
value += ' ';
elif key and value:
dict[key] = value
key = None
value = ''
else:
value += c
dict[key] = value
print dict