I want to aggregate data by hourly mean. Daily is very easy:
apply.daily(X2,mean)
Why is there no function for hourly? I tried
hr.means <- aggregate(X2, format(X2["timestamp"],"%Y-%m-%d %H"))
and got always error with trim argument. Is there an easier function similar to apply.daily? What if I want to aggregate the mean of 5 minutes. Data are values per minute:
"timestamp", value "2012-04-09 05:03:00",2 "2012-04-09 05:04:00",4 "2012-04-09 05:05:00",5 "2012-04-09 05:06:00",0 "2012-04-09 05:07:00",0 "2012-04-09 05:08:00",3 "2012-04-09 05:09:00",0 "2012-04-09 05:10:00",1
I am using xts and zoo.
try
period.apply(X2, endpoints(X2, "hours"), mean)
apply.daily is simply a wrapper for the above:
> apply.daily function (x, FUN, ...) { ep <- endpoints(x, "days") period.apply(x, ep, FUN, ...) }
hr.means <- aggregate(X2, format(time(X2),"%y-%m-%d %H"), mean)
This should work fine.
Answering part 2:
What if I want to aggregate the mean of 5 minutes?
As @eddit already mentioned in a comment above:
df <- read.table(header=TRUE, sep=",", stringsAsFactors=FALSE, text=" timestamp, value 2012-04-09 05:03:00,2 2012-04-09 05:04:00,4 2012-04-09 05:05:00,5 2012-04-09 05:06:00,0 2012-04-09 05:07:00,0 2012-04-09 05:08:00,3 2012-04-09 05:09:00,0 2012-04-09 05:10:00,1") X2 <- xts(df$value, as.POSIXct(df$timestamp)) X2.5min <- period.apply(X2, endpoints(X2, "minutes", 5), mean)
I get: 05:04:00 - 4; 05:09:00 - 5,... but maybe it is possible to set the first value to 05:00:00 and go on with 05:05:00 might be easier, if I am merging files later to have the same start and timestep.
Indeed:
> X2.5min [,1] 2012-04-09 05:04:00 3.0 2012-04-09 05:09:00 1.6 2012-04-09 05:10:00 1.0
Darren Cook over at Cross Validated faced the same issue and wrote function align.time.down:
align.time.down=function(x,n){index(x)=index(x)-n;align.time(x,n)}
That can be used to adjust the times down:
X2.5mindown <- align.time.down(X2.5min, 5 * 60) X2.5mindown [,1] 2012-04-09 05:00:00 3.0 2012-04-09 05:05:00 1.6 2012-04-09 05:10:00 1.0
