Pandas: How can I add two timestamp values?

Question

I am trying to add more than two timestamp values and I expect to see output in minutes/seconds. How can I add two timestamps? I basically want to do: '1995-07-01 00:00:01' + '1995-07-01 00:05:06' and see if total time>=60minutes. I tried this code: df['timestamp'][0]+df['timestamp'][1]. I referred this post but my timestamps are coming from dataframe. Head of my dataframe column looks like this:

0   1995-07-01 00:00:01
1   1995-07-01 00:00:06
2   1995-07-01 00:00:09
3   1995-07-01 00:00:09
4   1995-07-01 00:00:09
Name: timestamp, dtype: datetime64[ns]

I am getting this error: TypeError: unsupported operand type(s) for +: 'Timestamp' and 'Timestamp'

How do you expect to add two timestamps? If you add Tuesday and Wednesday, does that make Friday? Please show your expected output for given input. — Stephen Rauch, Apr 02 '17 at 00:53
I want to add two timestamps '1995-07-01 00:00:01' + '1995-07-01 00:59:59' to see the output like this 60minutes. — jubins, Apr 02 '17 at 01:00
yes, I have dataframe of IP addresses and timestamps. I want to know total duration for each IP address. I am trying to calculate this by adding all timestamp values for each unique IP address. — jubins, Apr 02 '17 at 01:04
You missed the important part of the question. Duration from when? — Stephen Rauch, Apr 02 '17 at 01:11
I am not trying to subtract. If you see head of my dataframe I am trying to add time values to calculate total duration. — jubins, Apr 02 '17 at 01:15
Sigh, I will go back to my first question. If you add Tuesday and Wednesday, does that make Friday? — Stephen Rauch, Apr 02 '17 at 01:20
No it won't make Friday. I'm trying my best to make it clear but I'm not sure what is missing to make you understand. I hope to get solution for this soon. I am stuck on it since 3 hours. — jubins, Apr 02 '17 at 01:29
A duration always requires a subtraction if you have a timestamp. Because a duration has a beginning and an end. You are trying to add Tuesday and Wednesday. — Stephen Rauch, Apr 02 '17 at 01:32

score 3 · Answer 1 · answered Apr 02 '17 at 02:29

The problem is that adding Timestamps makes no sense. What if they were on different days? What you want are the sum of Timedeltas. We can create Timedeltas by subtracting a common date from the whole series. Let's subtract the minimum date. Then sum up the Timedeltas. Let s be your series of Timestamps

s.sub(s.dt.date.min()).sum().total_seconds()

34.0

score 2 · Accepted Answer · answered Apr 02 '17 at 02:26

#Adding two timestamps is not supported and not logical
#Probably, you really want to add the time rather than the timestamp itself
#This is how to extract the time from the timestamp then summing it up

import datetime
import time

t = ['1995-07-01 00:00:01','1995-07-01 00:00:06','1995-07-01 00:00:09','1995-07-01 00:00:09','1995-07-01 00:00:09']
tSum = datetime.timedelta()
df = pd.DataFrame(t, columns=['timestamp'])
for i in range(len(df)):
    df['timestamp'][i] = datetime.datetime.strptime(df['timestamp'][i], "%Y-%m-%d %H:%M:%S").time()
    dt=df['timestamp'][i]
    (hr, mi, sec) = (dt.hour, dt.minute, dt.second)
    sum = datetime.timedelta(hours=int(hr), minutes=int(mi),seconds=int(sec))
    tSum += sum
if tSum.seconds >= 60*60:
    print("more than 1 hour")
else:
    print("less than 1 hour")

Pandas: How can I add two timestamp values?

2 Answers2