CODEWITHSUNDEEP
rtime-series
Yair Daon
Yair Daon

Reputation: 1123

Partition time series data to segments of a week increase

I have an irregularly sampled time series. I would like to partition it to segments where the time difference between consecutive entries is (say) a week. For example:

day     X
1       4
3       2
7       4
8       9
10      2
12      4
14      3
15      9
17      7
19      3
26      9

would be partitioned to: 

day     X
1       4
8       9
15      9

day     X
3       2
10      2
17      7

day     X 
12      4
19      3
26      9

day     X
7       4
14      3

I imagine I can brute force it with loops but I'm sure there are better ways since this seems like a fairly common task. Bonus points - instead of "week" use a range of days (say 6-8 days).

Upvotes: 1

Views: 98

Answers (1)

akrun
akrun

Reputation: 887541

We can use split to return a list of data.frames

split(df, df$day%%7)
#$`0`
#  day X
#3   7 4
#7  14 3

#$`1`
#  day X
#1   1 4
#4   8 9
#8  15 9

#$`3`
#  day X
#2   3 2
#5  10 2
#9  17 7

#$`5`
#   day X
#6   12 4
#10  19 3
#11  26 9

Upvotes: 2

Related Questions