using tapply to filter a variable

[Martyn Colins <martyncollins10@yahoo.com>]

 

I have a question related to the tapply function. If I have the following data.frame:

 

w <- data.frame(y1 =1:10, y2=as.factor(rep(LETTERS[1:5],2)), y3=as.factor(rep(c("level1", "level2"),5)), y4=c(rep(c("sub1", "sub2", "sub3"), 3), "sub3"))

 

and I want to sum w$y1 for w$y2 and w$y3 when w$y4 is equal to "sub1" - I tried this

tapply(w$y1, list(w$y2, w$y3, w$y4=="sub1"), sum)

 

However, I get 2 tables (FALSE and TRUE) like this

> tapply(w$y1, list(w$y2, w$y3, w$y4 == "sub1"), sum)

, , FALSE

level1 level2

A NA 6

B NA 2

C 3 8

D 9 NA

E 5 10

, , TRUE

level1 level2

A 1 NA

B 7 NA

C NA NA

D NA 4

E NA NA

 

How do I get the table for "sub1"  without performing the "sum" operation for when w$y4 is not equal to "sub1", i.e. I want the TRUE table without calculating the FALSE table as I'm performing a computation intensive operation?

 

Thanks in advance for any help.

 

Martyn 

---

Yahoo! Music Unlimited - Access over 1 million songs.
Try it free.