JIYIK CN >

We will demonstrate how to get the sum of Pandas groupby and sum. We will also look at pivotthe function to arrange the data in a nice table and how to define custom functions and apply them to DataFrame. We can also get the sum by using agg().

groupbyThe cumulative sum of

We can use groupbythe method to get the cumulative sum. For example, the following has date, fruit name and sales DataFrame:

import pandas as pd

df = pd.DataFrame(
    {
        "Date": ["08/09/2018", "10/09/2018", "08/09/2018", "10/09/2018"],
        "Fruit": [
            "Apple",
            "Apple",
            "Banana",
            "Banana",
        ],
        "Sale": [34, 12, 22, 27],
    }
)

If we want to calculate the cumulative sales total for each fruit on each date, we can do the following:

import pandas as pd

df = pd.DataFrame(
    {
        "Date": ["08/09/2018", "10/09/2018", "08/09/2018", "10/09/2018"],
        "Fruit": [
            "Apple",
            "Apple",
            "Banana",
            "Banana",
        ],
        "Sale": [34, 12, 22, 27],
    }
)
print(df.groupby(by=["Fruit", "Date"]).sum().groupby(level=[0]).cumsum())

Output:

Fruit  Date         Sale
Apple  08/09/2018    34
       10/09/2018    46
Banana 08/09/2018    22
       10/09/2018    49

pivot()Rearrange data in a nice table

pivot()The method can set the table's row and column properties. Let's change the above code and apply pivot()the method to rearrange the data in a beautiful table:

import pandas as pd

df = pd.DataFrame(
    {
        "Date": ["08/09/2018", "10/09/2018", "08/09/2018", "10/09/2018"],
        "Fruit": [
            "Apple",
            "Apple",
            "Banana",
            "Banana",
        ],
        "Sale": [34, 12, 22, 27],
    }
)
print(
    df.groupby(["Fruit", "Date"], as_index=False).sum().pivot("Fruit", "Date").fillna(0)
)

Output:

             Sale           
Date   08/09/2018 10/09/2018
Fruit                       
Apple          34         12
Banana         22         27

Applying functions to Pandasgroupby

We'll create a simple method to get seriesthe count of values ​​in a , or one-dimensional array, and use groupbyto get the total count for each value:

from pandas import *

d = {"series": Series(["1", "2", "1", "1", "4", "4", "5"])}
df = DataFrame(d)


def get_count(values):
    return len(values)


grouped_count = df.groupby("series").series.agg(get_count)
print(grouped_count)

After running the code, we will get the following output, which provides seriesthe number of occurrences of each value in .

Output:

series
1    3
2    1
4    2
5    1
Name: series, dtype: int64

agg()Get the sum of a column

We can use agg()to apply a sum operation on a column. Consider the following code:

import pandas as pd

df = pd.DataFrame(
    {
        "Date": ["08/09/2018", "10/09/2018", "08/09/2018", "10/09/2018"],
        "Fruit": [
            "Apple",
            "Apple",
            "Banana",
            "Banana",
        ],
        "Sale": [34, 12, 22, 27],
    }
)
print(df.groupby(["Fruit"])["Sale"].agg("sum"))

Output:

Fruit
Apple     46
Banana    49
Name: Sale, dtype: int64