I have a data frame that looks like this:
Timedate TotalSolar_MW
20 2013-06-01 04:45:00 13.0
21 2013-06-01 05:00:00 41.7
22 2013-06-01 05:15:00 81.8
23 2013-06-01 05:30:00 153.0
24 2013-06-01 05:45:00 270.7
25 2013-06-01 06:00:00 429.3
26 2013-06-01 06:15:00 535.4
"Timedate" is POSIXlt
, and "Total_Solar" is numeric
. The time steps are in 15 minute intervals from June 1, 0:00 to June 24, 24:00.
Now I want to aggregate
the quarter hourly data to hourly steps e.g. 2013-06-01 06:00:00 934.8MW (81.8MW + 153.0MW + 270.7MW + 429.3MW; from 05:15 to 06:00)
I tried this with:
Sum <-aggregate(Total_Solar_Gesamt$TotalSolar_MW,
list(as.POSIXlt(Total_Solar_Gesamt$Timedate)$hour), FUN=sum)
But it returns the aggregated hourly data of the whole data frame and gives me a new data.frame with 24 rows and the summed up MW for every hour.
How can I change the structure, only to reduce from a quarter hourly to an hourly interval? I tried a for loop but this also didn't work. Also subset
didn't work for me.
When working with time series, I suggest you work with xts
package for this, and for example hourly.apply
:
library(xts)
dat.xts <- xts(Total_Solar_Gesamt$TotalSolar_MW,
as.POSIXct(otal_Solar_Gesamt$Timedate))
hourly.apply(dat.xts,sum)
More general you can use period.apply
which is (lapply
equivalent) , for example to aggregate your data each 2 hours you can do the following:
ends <- endpoints(zoo.data,'hours',2)
period.apply(dat.xts,ends ,sum)