How do you take a string and then call a data frame with name = that string?

Question

The following code returns a string called "GLD".

CatItUp <- function(x){
    print(x)
}

CatItUp("GLD")

This code returns the tail of GLD prices. But obviously, because I've hard-coded GLD into the function.

IAmMoney <- function(x) {

    require("quantmod")

    getSymbols("GLD")

    tail(GLD)   

}

IAmMoney("GLD")

This does not return prices like the hard-coded version, but the "GLD" string like the CatItUp() example above. I don't know why.

IAmMoney <- function(x) {

    require("quantmod")

    getSymbols("x")

    tail(x) 

}

IAmMoney("GLD")

How can you pass 'GLD' to the quantmod::getSymbols function, inside the IAmMoney() function?

Answer 1

Aren't you simply overlooking the fact that getSymbols() has an option auto.assign ?

So you may want this instead:

R> library(quantmod)
R> silly <- function(sym) {
+     x <- getSymbols(sym, auto.assign=FALSE)
+     tail(x)
+ }
R> silly("IBM")
           IBM.Open IBM.High IBM.Low IBM.Close IBM.Volume IBM.Adjusted
2010-12-03   144.25   145.68  144.25    145.38    3710600       145.38
2010-12-06   144.54   145.87  144.52    144.99    3321800       144.99
2010-12-07   146.02   146.30  143.87    144.02    4828600       144.02
2010-12-08   144.35   145.65  143.84    144.98    4961400       144.98
2010-12-09   145.94   145.94  143.52    144.30    4405300       144.30
2010-12-10   144.88   144.95  143.73    144.82    3503800       144.82
R> silly("C")
           C.Open C.High C.Low C.Close   C.Volume C.Adjusted
2010-12-03   4.38   4.46  4.35    4.45  360277300       4.45
2010-12-06   4.45   4.50  4.43    4.45  445170000       4.45
2010-12-07   4.55   4.65  4.54    4.62 3265796000       4.62
2010-12-08   4.61   4.64  4.55    4.64  913820900       4.64
2010-12-09   4.68   4.71  4.64    4.69  731119000       4.69
2010-12-10   4.70   4.77  4.66    4.77  763156100       4.77
R> 

getSymbols() default behaviour of "I will stick it into your environment as a new variable" is more or less a design flaw and, as I recall, recognised as such.

And hence the behaviour can be altered by auto.assign.

Answer 2

Will tail(get(x)) work?

Answer 3

It's tricky because quantmod creates a data frame who's name is the same as the string you assign to x. So at first you need a string value then later you are calling a data frame by the name of x. This is exactly what do.call() is helpful for.

IAmMoney <- function(x) {
    require("quantmod")
    getSymbols(x)
    tail(get(x))
    # changed to get(x) per Ahala's input below. 
    # if you had many params you were passing, do.call()
    # might make more sense
}

IAmMoney("GLD")

Dirk pointed out that the use of the auto.assign=FALSE argument which means you can simply do this instead:

tail(getSymbols("GLD", auto.assign=FALSE))