pandas: efficiently adding the values of two (time) series-CodePudding

I have two (larger than the toy example) Series e.g.:

s1 = pd.read_json('{"count":{"1614470400000":4,"1617148800000":0,"1619740800000":0,"1622419200000":4,"1625011200000":4,"1627689600000":5,"1630368000000":0,"1632960000000":8,"1635638400000":2}}')['count']
s2 = pd.read_json('{"count":{"1625011200000":1}}')['count']

What would be an efficient way to add them? I can imagine creating a join, then adding ignoring nans, but I think this should be slower than some operation that would merge sort them?

Here is the desired result for the toy example:

CodePudding user response：

Below works but I really don't know how you are adding june 30 2019 and june 30 2021 together to get 5. I had to change 2019 to 2021 to make it work.

s3 = pd.concat([s1,s2])
s3.groupby(level=0).sum()

CodePudding user response：

Try with Series.add(...) using the parameter fill_value=0

out = s1.add(s2,fill_value=0)
            count
2019-06-30    1.0
2021-02-28    4.0
2021-03-31    0.0
2021-04-30    0.0
2021-05-31    4.0
2021-06-30    4.0
2021-07-31    5.0
2021-08-31    0.0
2021-09-30    8.0
2021-10-31    2.0