I have a list of lists looking like this:
['user1', time, cpus, mem]
['user1', time, cpus, mem]
['user2', time, cpus, mem]
['user3', time, cpus, mem]
and so on..
I need just one list per user with time, cpus and mem added together. I've tried a few things but i cant make it work.
A sweet, Pythonic and concise way to do that would be:
from collections import defaultdict
l = [
['user1', 0, 1, 0],
['user2', 2, 2, 2],
['user3', 2, 2, 1],
['user3', 1, 1, 2],
['user1', 1, 0, 1],
]
merged = defaultdict(lambda: [0, 0, 0])
for user, *values in l:
merged[user] = [sum(i) for i in zip(values, merged[user])]
Output:
In : merged
Out:
defaultdict(<function __main__.<lambda>>,
{'user1': [1, 1, 1], 'user2': [2, 2, 2], 'user3': [3, 3, 3]})
That uses a defaultdict with a list of length 3 as its default value. The relevant user's values are updated for each element in the list.
You can do this using a dictionary in order to group list items by user.
Then just use zip function to find out the sum for every feature from time, cpus, mem list.
mylist = [['user1', 1, 6, 8],
['user1', 2, 7, 9],
['user2', 3, 7, 5],
['user3', 4, 7, 3]]
dict = {}
for elem in mylist:
if elem[0] not in dict:
dict[elem[0]] = []
dict[elem[0]].append(elem[1:])
for key in dict:
dict[key] = [sum(i) for i in zip(*dict[key])]
Output
In[1]: dict
Out[1]: {'user1': [3, 13, 17], 'user2': [3, 7, 5], 'user3': [4, 7, 3]}
dict as a variable name since it is a built-in function docs.python.org/3/library/functions.html .What about putting the list of lists into a pandas.DataFrame.
Then you could do grouping by users and apply aggregate functions like sum().
df = pd.DataFrame()
df['users'] = ['user1', 'user1','user2', 'user3']
df['time'] = [1,2,3,4]
df['mem'] = [128, 256, 1024, 2048]
df['cpu'] = [4, 2, 16, 32]
In: df
out:
users time cpu mem
0 user1 1 4 128
1 user1 2 2 256
2 user2 3 16 1024
3 user3 4 32 2048
In: df.groupby('users').sum()
Out:
time cpu mem
users
user1 3 6 384
user2 3 16 1024
user3 4 32 2048
def gropulistsinlist(data,columns):
ll=[[]]
#ll.append([[3,4,5,6,7,7]])
#columns=[4,5]
#data=[[3,4,5,6,7,8],[3,4,5,6,7,7],[3,4,5,6,7,9],[3,4,5,6,7,8]]
#data=[[3,4,5,6,7,7],[3,4,5,6,7,8],[3,4,5,6,7,7],[3,4,5,6,7,9],[3,4,5,6,7,8]]
for d in data:
addnew=0
addwhat=0
addwhere=0
for index,l in enumerate(ll):
if len(l)and([l[0][ii] for ii in columns]== [d[iii] for iii in columns]):
addwhere=index
addwhat=d
addnew=0
#print(ll)
#print()
break
addnew=1
addwhat=d
if(addnew):
ll.append([addwhat])
else:
ll[index].append(addwhat)
ll.pop(0)
return ll
ll will be
[
[[3, 4, 5, 6, 7, 7], [3, 4, 5, 6, 7, 7]],
[[3, 4, 5, 6, 7, 8],[3, 4, 5, 6, 7, 8]],
[[3, 4, 5, 6, 7, 9]]]
dictwith users as keys ?