Skip to content

More effective calculation for loop

3 messages · Wolfgang Wu, John Kane, Joshua Ulrich

#
I am getting confused in how to use R effectively. Below is the example that I 
would like to optimize, which is basically a loop. In its current version it 
takes  about 20 seconds to run. I have looked at the apply function and but 
didn't seem to get it to work. I am sure this is relatively easy to show me a 
better version for someone who has done this before. For some reason I seem to 
have trouble getting my head around this. 

#----------------------------------
#Create the example data
dDates <- seq(from=as.Date("1990-01-01"), by="day", length.out=10000);
xtsHold <- xts(rep(0, times=10000), order.by=dDates);
xtsBuy <- xts(sample(0:1,10000,replace=T), order.by=dDates);
xtsSell <- xts(sample(0:1,10000,replace=T), order.by=dDates);
print(Sys.time())

#Run the inefficient calculation
for (i in 2:length(xtsHold)){
xtsHold[i] <- xtsHold[i-1] + xtsBuy[i-1] - xtsSell[i-1];
}

print(Sys.time())#----------------------------------


Thank you very much in advance!

Best regards, 

Wolfgang Wu
#
What is xts?  What package?
--- On Fri, 5/13/11, Wolfgang Wu <wobwu22 at yahoo.de> wrote:

            
#
On Fri, May 13, 2011 at 4:32 AM, Wolfgang Wu <wobwu22 at yahoo.de> wrote:
Use a vectorized solution instead:

xtsHold2 <- lag(xtsBuy)-lag(xtsSell)
xtsHold2[1] <- 0
xtsHold2 <- cumsum(xtsHold2)
identical(xtsHold, xtsHold2)
--
Joshua Ulrich  |  FOSS Trading: www.fosstrading.com