Tapply Approach

[ledzep]

Hi,

I use ddply a lot which works like a charm for small datasets. However, when it comes to large datasets, it falters [I was waiting forever for my ddply progress bar to reach 100%].

Dason suggested the use of tapply. I was very impressed with how quick tapply was. However, I couldn't quite figure out how to use merge the lists back to the parent data frame.

# Example data
test<-structure(list(id = c("A", "A", "A", "A", "A", "A", "A", 
"A", "B", "B", "B", "B", "B", "B", "B", "B", "C", "C", "C", "C", 
"C", "C", "C", "C"), date = c("2011-01-11", "2011-01-12", "2011-01-13", 
"2011-01-14", "2011-01-18", "2011-01-25", "2011-02-01", "2011-02-08", 
"2011-01-11", "2011-01-12", "2011-01-13", "2011-01-14", "2011-01-18", 
"2011-01-25", "2011-02-01", "2011-02-08", "2011-01-11", "2011-01-12", 
"2011-01-13", "2011-01-14", "2011-01-18", "2011-01-25", "2011-02-01", 
"2011-02-08")), .Names = c("id", "date"), row.names = c("6", 
"5", "3", "1", "2", "8", "7", "4", "12", "15", "11", "16", "14", 
"9", "13", "10", "17", "19", "20", "22", "21", "18", "23", "24"
), class = "data.frame")

test$date<-as.Date(test$date) # declare date

Now use tapply.

# Work out successive difference between rows
Diff_calc<- function(x){
		(c(NA, diff(x))) # or replace the NAs as 0
		}



# test$diff<-tapply(test$date, test$id, Diff_calc) # doesn't work very well

# Instead use
diff<-tapply(test$date, test$id, Diff_calc)
> diff
$A
[1] NA  1  1  1  4  7  7  7

$B
[1] NA  1  1  1  4  7  7  7

$C
[1] NA  1  1  1  4  7  7  7

How would I bind these lists back to my parent date frame?

Many Thanks

[Jake]

test$diff <- unlist(diff)