How to merge and sum duplicates in python?

Question

I have a numpy array:

   [2016-10-20 00:00:00, 6, 17],
   [2016-10-20 00:00:00, 0, 21],
   [2017-09-11 00:00:00, 7, 22],
   [2017-09-11 00:00:00, 5, 30],
   [2017-09-11 00:00:00, 2, 40]

Is there an easy way to merge the the rows with the same dates and sum up the values belonging to the same date?

For the example above, It should look like this:

   [2016-10-20 00:00:00, 6, 38],
   [2017-09-11 00:00:00, 14, 92]

Similar question http://stackoverflow.com/questions/4373631/sum-array-by-number-in-numpy — Xiongbing Jin, Mar 12 '16 at 18:11
That doesn't look like a valid NumPy array format. Also, are the dates sorted in that sequence? — Divakar, Mar 12 '16 at 18:16

score 4 · Accepted Answer · answered Mar 12 '16 at 18:58

I suggest using pandas for this:

import pandas as pd

d = pd.DataFrame([
    ['2016-10-20 00:00:00', 6, 17],
    ['2016-10-20 00:00:00', 0, 21],
    ['2017-09-11 00:00:00', 7, 22],
    ['2017-09-11 00:00:00', 5, 30],
    ['2017-09-11 00:00:00', 2, 40]
])

d.groupby([0]).aggregate(sum)

#                       1   2
# 0
# 2016-10-20 00:00:00   6  38
# 2017-09-11 00:00:00  14  92

Note that pandas.DataFrame accepts a numpy array as input.

How to merge and sum duplicates in python?

1 Answers1