CODEWITHSUNDEEP
rzooperformanceanalyticsrollapply
Jairaj Gupta
Jairaj Gupta

Reputation: 367

Using apply.rolling function over multiple columns

In zoo package, while using rollapply function, we get an logical option 'by.column = TRUE/FALSE'. If TRUE, FUN is applied to each column separately.

However, in apply.rolling function of PerformanceAnalytics package I don't see any such option. While trying to the following:

require(PerformanceAnalytics)

data(managers)

apply.rolling(managers[,1:10,drop=FALSE], FUN="mean", width=36)

It give me calculation of only first column. Can someone please tell me how can I run this calculation over all the columns in the dataset.

Upvotes: 4

Views: 3264

Answers (4)

onlyphantom
onlyphantom

Reputation: 9623

You can combine lapply() with rollapply() from the zoo package to perform your rolling function over multiple columns. In the following code, I select all columns except the first (select(-1)) and perform my rolling function over the remaining columns.

In dplyr() code:

m3v %>% 
  select(-1) %>% 
  lapply(function(x){rollapply(x, 
                           width=14, 
                           FUN=function(x){sum(x, na.rm=T)},
                           partial = T, fill = 0
                           )}) %>% 
  as.data.frame()

Upvotes: 0

lukehawk
lukehawk

Reputation: 1493

In the help page for apply.rolling it says:

Details

Wrapper function for rollapply to hide some of the complexity of managing **single-column zoo objects**.

I think that means the intent is to not use this on objects with multiple columns. It is a simplified wrapper which allows you to not have to type out all the various optional parameters for rollapply().

Upvotes: 0

Davis Vaughan
Davis Vaughan

Reputation: 2960

Is there something wrong with just using rollapply()? apply.rolling() is just a wrapper over it.

rollapply(managers, FUN = mean, width = 36, by.column = TRUE)

Upvotes: 4

Lstat
Lstat

Reputation: 1460

You can apply function apply.rolling on xts object:

apply(managers, 2, function(x){apply.rolling(x, FUN="mean", width=36)})

Upvotes: 1

Related Questions