Python sum values of list of dictionaries if two other key value pairs match

Question

I have a list of dictionaries of the following form:

lst = [{"Name":'Nick','Hour':0,'Value':2.75},
       {"Name":'Sam','Hour':1,'Value':7.0},
       {"Name":'Nick','Hour':0,'Value':2.21},
       {'Name':'Val',"Hour":1,'Value':10.1},
       {'Name':'Nick','Hour':1,'Value':2.1},  
       {'Name':'Val',"Hour":1,'Value':11},]  

I want to be able to sum all values for a name for a particular hour, eg if Name == Nick and Hour == 0 , I want value to give me the sum of all values meeting the condition. 2.75 + 2.21 , according to the piece above.

I have already tried the following but it doesn't help me out with both conditions.

finalList = collections.defaultdict(float)
for info in lst:
    finalList[info['Name']] += info['Value']
finalList = [{'Name': c, 'Value': finalList[c]} for c in finalList]  

This sums up all the values for a particular Name , not checking if the Hour was the same. How can I incorporate that condition into my code as well?

My expected output :

finalList = [{"Name":'Nick','Hour':0,'Value':4.96},
       {"Name":'Sam','Hour':1,'Value':7.0},
       {'Name':'Val',"Hour":1,'Value':21.1},
       {'Name':'Nick','Hour':1,'Value':2.1}...]  

Answer 1

consider using pandas module - it's very comfortable for such data sets:

import pandas as pd

In [109]: lst
Out[109]:
[{'Hour': 0, 'Name': 'Nick', 'Value': 2.75},
 {'Hour': 1, 'Name': 'Sam', 'Value': 7.0},
 {'Hour': 0, 'Name': 'Nick', 'Value': 2.21},
 {'Hour': 1, 'Name': 'Val', 'Value': 10.1},
 {'Hour': 1, 'Name': 'Nick', 'Value': 2.1}]

In [110]: df = pd.DataFrame(lst)

In [111]: df
Out[111]:
   Hour  Name  Value
0     0  Nick   2.75
1     1   Sam   7.00
2     0  Nick   2.21
3     1   Val  10.10
4     1  Nick   2.10

In [123]: df.groupby(['Name','Hour']).sum().reset_index()
Out[123]:
   Name  Hour  Value
0  Nick     0   4.96
1  Nick     1   2.10
2   Sam     1   7.00
3   Val     1  10.10

export it to CSV:

df.groupby(['Name','Hour']).sum().reset_index().to_csv('/path/to/file.csv', index=False)

result:

Name,Hour,Value
Nick,0,4.96
Nick,1,2.1
Sam,1,7.0
Val,1,10.1

if you want to have it as a dictionary:

In [125]: df.groupby(['Name','Hour']).sum().reset_index().to_dict('r')
Out[125]:
[{'Hour': 0, 'Name': 'Nick', 'Value': 4.96},
 {'Hour': 1, 'Name': 'Nick', 'Value': 2.1},
 {'Hour': 1, 'Name': 'Sam', 'Value': 7.0},
 {'Hour': 1, 'Name': 'Val', 'Value': 10.1}]

you can do many fancy things using pandas:

In [112]: df.loc[(df.Name == 'Nick') & (df.Hour == 0), 'Value'].sum()
Out[112]: 4.96


In [121]: df.groupby('Name')['Value'].agg(['sum','mean'])
Out[121]:
        sum       mean
Name
Nick   7.06   2.353333
Sam    7.00   7.000000
Val   10.10  10.100000

Answer 2

[{'Name':name, 'Hour':hour, 'Value': sum(d['Value'] for d in lst if d['Name']==name and d['Hour']==hour)} for hour in hours for name in names]

if you don't already have all names and hours in lists (or sets) you can get them like so:

names = {d['Name'] for d in lst}
hours= {d['Hour'] for d in lst}

Answer 3

You can use any (hashable) object as a key for a python dictionary, so just use a tuple containing Name and Hour as the key:

from collections import defaultdict
d = defaultdict(float)
for item in lst:
    d[(item['Name'], item['Hour'])] += item['Value']