split xts object by index (date)

Question

I have an xts object that has intraday data:

head(stocks[,1])
                    SMH.close
2009-01-02 09:31:00     17.66
2009-01-02 09:32:00     17.66
2009-01-02 09:33:00     17.64
2009-01-02 09:34:00     17.60
2009-01-02 09:35:00     17.58
2009-01-02 09:36:00     17.63

I want to perform various analytics on intraday data but operations should not cross day boundaries. So what I want to do is split the data by date (ignoring the time). As such I extracted the index and saved unique date values by:

y <- index(stocks)
x <- strptime(y, format="%Y-%m-%d")
uniquedates <- unique(x)

Now I would like to do something similar to the example in ?split

> g <- airquality$Month
> l <- split(airquality, g)

Here the airquality data is split into a list object by the value of the Month column. I'm not sure how to do something similar as the date is the index in my case and not a data column. I tried but get an error.

> split(stocks, uniquedates)
Error in args[[i]] : subscript out of bounds

Perhaps there is a cleaner way of achieving what I want to do. I would greatly appreciate your help.

IRTFM · Accepted Answer · 2011-05-15 02:06:43Z

3

You should skip the unique() step. Just use split.xts with an f="days" argument.

data(sample_matrix)
sample.xts <- as.xts(sample_matrix, descr='my new xts object')
split.xts(sample.xts, f="days")
[[1]]
               Open     High      Low    Close
2007-01-02 50.03978 50.11778 49.95041 50.11778

[[2]]
              Open     High     Low    Close
2007-01-03 50.2305 50.42188 50.2305 50.39767

[[3]]
               Open     High      Low    Close
2007-01-04 50.42096 50.42096 50.26414 50.33236

[[4]]
snipped

edited May 15, 2011 at 2:06

answered May 15, 2011 at 1:23

IRTFM

264k22 gold badges381 silver badges503 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Joshua Ulrich Over a year ago

+1 but split is a generic, so you don't need to explicitly call split.xts. split(sample.xts, "days") works.

IRTFM Over a year ago

Yes. I knew that but the arguments to split.xts are different than for split.default, so I decided to emphasize that difference.

Oleg Melnikov Over a year ago

Also, note that split() returns zoo objects, while split.xts returns xts/zoo objects.

Joshua Ulrich · Accepted Answer · 2011-05-14 23:35:18Z

0

Perhaps you're looking for apply.daily (or period.apply in general)?

answered May 14, 2011 at 23:35

Joshua Ulrich

177k33 gold badges357 silver badges429 bronze badges

2 Comments

codingknob Over a year ago

The period.apply functions will definitely come in use but I also want to compute various statistics - some rolling and others probability based - at every time increment X intraday. For example I want to know how many stocks are up / down at time = X * n (where n is an integer). As a result I think I need to split the data by day but perhaps there is a cleaner way preparing the data.

Joshua Ulrich Over a year ago

We could suggest the best alternatives if you gave a more complete example of what you're trying to do.

Collectives™ on Stack Overflow

split xts object by index (date)

2 Answers 2

3 Comments

2 Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

2 Comments

Related