Is there a fast way to search for variables in R?

Question

In Stata the lookfor command offers a quick way to search for variables in a dataset (and it searches both the variable names and labels). So lookfor education quickly finds you variables related to education. Is there an equivalent shortcut function in R?

Avouch

Voting to migrate to stackoverflow but you can combine the which() command with the names() command for this if you're working with a data frame, or colnames() if you're working with a matrix – Comanche

Answer 1

You can simply grep the data.frame for the information necessary. Then you'll get much more information than simply the list of names of variables for which somebody is matched. You can also use regular expressions, thus enhancing your search capabilities. Here is the example of a function which does what you want (works with data.frame only):

lookfor <- 
function (pattern, data, ...) 
{
    l <- lapply(data, function(x, ...) grep(pattern, x, ...))
    res <- rep(FALSE, ncol(data))
    res[grep(pattern, names(data), ...)] <- TRUE
    res <- sapply(l, length) > 0 | res
    names(res) <- names(data)
    names(res)[res]
}

First I grep each column, then I grep the column names. Then I keep only information whether grep matched anything and record it for each column separately. Instead of ... you can pass any arguments to grep. If you omit it, this function will do a simple string matching.

Here is an example:

> dt<- data.frame(y=1:10,x=letters[1:10],a=rnorm(10))
> lookfor("a",dt) 
[1] "x" "a"

Oliguria

Answer 2

How about this as a oneliner which I run at the start of a session:

lkf <- function(d,p) names(d)[grep(p,names(d))]

where d is the name of your data.frame and p is the pattern.

So

d <- data.frame(a=letters[1:10],b=1:10,c=month.name[1:10])
lkf(d,'c')
# [1] "c"

And here's a version that doesn't require you to quote the variable names

lookfor <- function(string_to_find, data){
    # Extract the arguments and force conversion to string
    pars <- as.list(match.call()[-1])
    data.name <- as.character(pars$data)
    var <- as.character(pars$string_to_find)

    # Regular expression search through names
    result <- names(data)[grep(var, names(data))]

    if(length(result) == 0) {
        warning(paste(var, "not found in", data.name))
        return(NULL)
    }
    else {
        return(result)
    }
}

Delatorre

Answer 3

If you just need to search though the list of variables to find the one you are looking for then it is possible to use the code completion function in RStudio (v0.99 onwards). Simply start typing and you will get a list of possible matches. So in your case type education$ and a list of variables contained in the data frame will appear. Scroll though these and select the one you want.

Nurseryman