Question

I couldn’t think how to title this one well, what I have is a vector of objects generated by a function (specifically midPci) of this form:

$conf.int
[1] 0.4726 0.6466
attr(,"conf.level")
[1] 0.95

$conf.int
[1] 0.1181 0.2566
attr(,"conf.level")
[1] 0.95

What I want to do is strip out the two values of $conf.int into vectors giving the upper and lower confidence intervals. Obviously I can do this with a loop, but I figured there was probably a better way to do this?

CIs <- structure(list(conf.int = structure(c(0.4696, 0.6501), conf.level = 0.95), 
 conf.int = structure(c(0.5266, 0.7081), conf.level = 0.95), conf.int = structure(c(0.4441, 
 0.6196), conf.level = 0.95), conf.int = structure(c(0.4181, 0.5891), conf.level = 0.95), 
 conf.int = structure(c(0.4726, 0.6466), conf.level = 0.95), conf.int = structure(c(0.1181, 
 0.2566), conf.level = 0.95), conf.int = structure(c(0, 0.000748652688017826), conf.level = 
 0.95)), .Names = c("conf.int", "conf.int", "conf.int", "conf.int", "conf.int", "conf.int", 
 "conf.int" ))

In case it’s not clear what I mean, I’ve done this:

# Calculate CIs
CIs <- mapply(midPci, k, n, conf.level=0.95)

And I can get the individual values I’m after out by doing CIs[1]$conf.int[1], etc., but I want to get the whole lot out in one go. Can I do this?

Answer 1

This produces a matrix:

do.call(rbind, CIs)

           [,1]         [,2]
conf.int 0.4696 0.6501000000
conf.int 0.5266 0.7081000000
conf.int 0.4441 0.6196000000
conf.int 0.4181 0.5891000000
conf.int 0.4726 0.6466000000
conf.int 0.1181 0.2566000000
conf.int 0.0000 0.0007486527

This also produces a matrix, although it is the ‘wide version’. The c function applied to a vector just returns itself , and you could have used as.vector or I and gotten the same result. It is really the function simplify2array(CIs) that is called by sapply that is doing the work:

 sapply(CIs, "c")
     conf.int conf.int conf.int conf.int conf.int conf.int     conf.int
[1,]   0.4696   0.5266   0.4441   0.4181   0.4726   0.1181 0.0000000000
[2,]   0.6501   0.7081   0.6196   0.5891   0.6466   0.2566 0.0007486527

And had we used do.call(cbind, CIs) we would have gotten the ‘wide version’ as well.