[R] Creating Functions in R

[Sarah Goslee]

Hi,


On Thu, Jul 24, 2014 at 9:35 AM, Pavneet Arora
<pavneet.arora at uk.rsagroup.com> wrote:
> Hello Guys
> I am new at writing Functions in R, and as a result am struggling with it.
> I am trying to use Google & other resources, but it's hard to find
> solutions when you don't know what to look for.

How about the introduction to R that comes with your installation?
It's got a section on writing
functions, and some other useful information that you seem to not have
learned yet.

> I have the following small dataset
>> dput(sub)
> structure(list(week = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,
> 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28,
> 29, 30), value = c(9.45, 7.99, 9.29, 11.66, 12.16, 10.18, 8.04,
> 11.46, 9.2, 10.34, 9.03, 11.47, 10.51, 9.4, 10.08, 9.37, 10.62,
> 10.31, 10, 13, 10.9, 9.33, 12.29, 11.5, 10.6, 11.08, 10.38, 11.62,
> 11.31, 10.52)), .Names = c("week", "value"), row.names = c(NA,
> -30L), class = "data.frame")
>
> I want to take each of the value and subtract from a target {in this case
> its 10}.

Thank you for providing data with dput()!

There are a bunch of things wrong with your function, starting with
the lack of need for a function.

If I understand your description correctly, what you actually want is:

sub$deviation <- sub$value - 10

But for educational purposes, here goes:

> This is what I have written in my function so far:
> vmask <- function(data,target){
> for(k in 1:length(data))

this actually loops through the COLUMNS of data, so first you're subtracting
target from week, then from value

> deviation <- data[k]- target

but coincidentally it gives you what you thought you were getting,
because you're overwriting deviation with each value of k, so the week
-target column is never saved. It's a really good idea to explicitly
mark the loop with { } too, to reduce confusion.

> dev <- return(data.frame(cbind(data,deviation)))

Hm. I don't know what you're trying to do with return() here, and
using both data.frame() and cbind() is superfluous. It isn't always
necessary, but I find it useful to explicitly name the columns of your
data frame when you create it, which gives

dev <- data.frame(data, deviation = deviation))

> return(dev)

The last item of a function is what's returned, so all you really need here is

dev

> }
> vmask(sub,10)
> View(dev)

dev only exists within the scope of the function. But you didn't
assign the return value of the function to anything. If you assign it
to an object named dev, then dev will exist in the global environment:

dev <- vmask(sub, 10)


> But when I run this I get the results as expected. But I expected the new
> coloumn to be called "deviation", whereas R just calls in "value.1". How
> can I fix this?
> Also I was hoping to see this new dataset with columns "week", "value",
> and now "deviation" when I use "View(dev) - but it comes up with error
> 'dev not found'. How can i fix this? Also is there anyway instead of me
> making a new dataset called "dev" with the 3 columns, I can just re-use my
> original dataset "sub" and give me all the 3 new columns?
>
> The next step I want to do is to perform a cumulative sum. So looking at
> the results, I want a new coloumn in existing dataset (or new dataset),
> which will now have 4 columns. The 4th column I want to be called "CuSum".
> So the first row of Cusum will be "-0.55", the second = "-0.55+(-2.01)"
> which will give me "-2.56" and so on forth.
>
> How can I do this in R using a function? Please help

You don't need a function. Just add the cumulative sum as a new column.

sub$Cusum <- cumsum(sub$deviation)


Sarah
-- 
Sarah Goslee
http://www.functionaldiversity.org