CODEWITHSUNDEEP
rpaste
Ken
Ken

Reputation: 13

Using R to access a structure dynamically with pasted commands?

I'm looking into comparing a set of lists (containing only strings) and seeing if their contents exist in a main list called "universe". The problem is I don't know upfront how many lists I will be looking at, could be from one to 20 or more - these are created dynamically and this works OK. The single line of code below works fine but defeats my objective. To work with the code below then I'll have to create a switch with as many options as potential lists (group) which is messy. 

allSE <- universe[group[[1]] & group[[2]] & group[[3]] ]

The two lines code below create an identical string to the commands in the square brackets above but simply doesn't work. It returns a single "NA" when I know there are 20 odd matches.

xnam <- paste0("group[[", 1:3,"]]")

allSE <- universe[paste(xnam, collapse= " & ")]

I've tried the formula command but that insists on a ~y since its for regression of course and not suitable - So how can I convince R it has commands between the square brackets and not a string for the command line in bold?

An example

universe <- c("ted","sara","fred","billy")

group1 <- as.logical(c("TRUE","TRUE","TRUE","TRUE"))

group2 <- as.logical(c("FALSE","TRUE","FALSE","TRUE"))

group3 <- as.logical(c("FALSE","TRUE","TRUE","TRUE"))

group <- list(group1, group2, group3)

universe[group[[1]] & group[[2]]  & group[[3]]]

Expected results

should give sara and billy as these two appear in all three lists

Upvotes: 0

Views: 43

Answers (1)

akrun
akrun

Reputation: 887611

We can use Reduce with & to compare the corresponding elements in the list that are all TRUE and it can be used to subset the 'universe'

universe[Reduce(`&`, group)]
#[1] "sara"  "billy"

Or another option is to rbind the list elements and do a colSums to check whether the number of TRUE elements are equal to the length of list

universe[colSums(do.call(rbind, group))==length(group)]

Upvotes: 2

Related Questions