CODEWITHSUNDEEP
pythonpandassum
Xavier M
Xavier M

Reputation: 549

How to sum similar pandas rows

How to sum the pandas rows from data_in in order to get panda rows similar to the lines of data_out?

data_in = [
    { 'col-a':'a1', 'col-b':'b1', 'col-z':'z1', 'value':1},
    { 'col-a':'a1', 'col-b':'b1', 'col-z':'z1', 'value':2},
    { 'col-a':'a2', 'col-b':'b2', 'col-z':'z2', 'value':10},
    { 'col-a':'a2', 'col-b':'b2', 'col-z':'z2', 'value':20}
]

df = pd.DataFrame(data_in)

# which operation to apply on df to get rows like in data_out?
# ...

data_out = [
    { 'col-a':'a1', 'col-b':'b1', 'col-z':'z1', 'value':3},
    { 'col-a':'a2', 'col-b':'b2', 'col-z':'z2', 'value':30}
]

Upvotes: 0

Views: 190

Answers (2)

Angelo
Angelo

Reputation: 655

This is the line you are looking for:

df.groupby(["col-a", "col-b", "col-z"])["value"].sum()

Upvotes: 1

Erfan
Erfan

Reputation: 42916

This is an aggregating problem. You can use .groupby in pandas and while aggregating you can get the sum of these groups with .value.sum()

df_out = df.groupby(['col-a', 'col-b', 'col-z']).value.sum().reset_index()
print(df_out)

  col-a col-b col-z  value
0    a1    b1    z1      3
1    a2    b2    z2     30

Upvotes: 2

Related Questions