Python Min/Max Dates with Groupby

Question

I am trying to get the start and end dates for particular price. In my example, the price is $3 for few days, then increases to $4 for few days and back to $3 later on.

import pandas as pd

df = pd.DataFrame([
    {"Price":3,"ds":"2017-01-01"},
    {"Price":3,"ds":"2017-01-02"},
    {"Price":3,"ds":"2017-01-03"},
    {"Price":3,"ds":"2017-01-04"},
    {"Price":3,"ds":"2017-01-05"},
    {"Price":4,"ds":"2017-01-06"},
    {"Price":4,"ds":"2017-01-07"},
    {"Price":4,"ds":"2017-01-08"},
    {"Price":4,"ds":"2017-01-09"},
    {"Price":3,"ds":"2017-01-10"},
    {"Price":3,"ds":"2017-01-11"},
    {"Price":3,"ds":"2017-01-12"},
    {"Price":3,"ds":"2017-01-13"},
    {"Price":3,"ds":"2017-01-14"}
    ])

start = df.groupby(["Price"])["ds"].min().reset_index()
end = df.groupby(["Price"])["ds"].max().reset_index()
df2 = start.merge(end, how = 'left', on = ["Price"])
df2.rename(columns = {"ds_x":"start_dt", "ds_y":"end_dt"},inplace=True)

Output:

   Price start_dt    end_dt
     3  2017-01-01  2017-01-14
     4  2017-01-06  2017-01-09

Desired Output:

   Price start_dt    end_dt
     3  2017-01-01  2017-01-05
     4  2017-01-06  2017-01-09
     3  2017-01-10  2017-01-14

Answer 1

Starting from your original dataframe, you can use a helper column with series.shift to compare the next row and use it for grouping, then groupby and agg with min and max, rename and reset the index:

s = df['Price'].ne(df['Price'].shift()).cumsum()

d = {"min":"start_dt", "max":"end_dt"}
out = (df.groupby([s,'Price'])['ds'].agg(['min','max']).rename(columns=d)
                                             .droplevel(0).reset_index())

---

print(out)

   Price    start_dt      end_dt
0      3  2017-01-01  2017-01-05
1      4  2017-01-06  2017-01-09
2      3  2017-01-10  2017-01-14