R help - May 2008 - Accesing data frame members from within functions

I am writing a simple R program to execute a t-test repeatedly on data
contained in a data frame.  My data looks like this:



Category                      Value1             Value2

1                                  .5                     .8

1                                  .3                     .9

. . .                               . . .                   . . .

2                                  1.4                   1.3

2                                  1.3                   1.3

. . .                               . . .                   . . .

15                                .2                     .3

15                                .5                     .1





So in all there are 15 categories, and each category contains two sets of
observations which I want to compare.  I only want to compare Value1 and
Value2 within each category, but I need to do it 15 times (once for each
category), so I wanted to write an R function to make it easier.



Right now I am using a for() loop to do the comparison.  My loop looks like
this:



for(i in 1:21)

{

            x <- t.test(Value1[Category == i], Value2[Category == i])

            y <- c(y, x$p.value)

}



The loop runs and everything is working well.  However, I am not sure how to
translate this code into a function.  In particular, I'm not sure how to
write a function that passes a data frame ds (containing Category, Value1,
and Value2 as members) as an argument, and then accessing these members
within the body of the function.  I've tried the following:



repeated_test <- function(data)

{

            for(i in 1:21)

            {

              x <- t.test(ds$Value1[ds$Category == i], ds$Value2[ds$Category
== i])

              y <- c(y, x$p.value)

            }



This will run, but only if the members of the data frame I am passing as an
argument are in fact named Value1, Value2, and Category.  This is fine for
now, but in the future I will have to run this function on data where I
cannot be sure this is the case.  Rather than change the member names by
hand, I would like to make the function generic to work with any data frame.
How do I do this?  Or is there a better way to do this without the for()
loop (for example, using apply())?

	[[alternative HTML version deleted]]

Try:

foo <- function(data, ...)
{
res <- unlist(lapply(split(data, data$Category),
        function(.x)t.test(.x$Value1, .x$Value2)$p.value))
test <- merge(data, as.data.frame(res), by.x="Category", by.y = 0)
return(test)
}

x <- data.frame(Category = rep(1:15, each = 10), Value1 = rnorm(150), Value2
= rnorm(150))
foo(x)

On Fri, May 2, 2008 at 11:19 AM, David Schwab <dvschwab46225@gmail.com>
wrote:
> I am writing a simple R program to execute a t-test repeatedly on data
> contained in a data frame.  My data looks like this:
>
>
>
> Category                      Value1             Value2
>
> 1                                  .5                     .8
>
> 1                                  .3                     .9
>
> . . .                               . . .                   . . .
>
> 2                                  1.4                   1.3
>
> 2                                  1.3                   1.3
>
> . . .                               . . .                   . . .
>
> 15                                .2                     .3
>
> 15                                .5                     .1
>
>
>
>
>
> So in all there are 15 categories, and each category contains two sets of
> observations which I want to compare.  I only want to compare Value1 and
> Value2 within each category, but I need to do it 15 times (once for each
> category), so I wanted to write an R function to make it easier.
>
>
>
> Right now I am using a for() loop to do the comparison.  My loop looks
> like
> this:
>
>
>
> for(i in 1:21)
>
> {
>
>            x <- t.test(Value1[Category == i], Value2[Category == i])
>
>            y <- c(y, x$p.value)
>
> }
>
>
>
> The loop runs and everything is working well.  However, I am not sure how
> to
> translate this code into a function.  In particular, I'm not sure how
to
> write a function that passes a data frame ds (containing Category, Value1,
> and Value2 as members) as an argument, and then accessing these members
> within the body of the function.  I've tried the following:
>
>
>
> repeated_test <- function(data)
>
> {
>
>            for(i in 1:21)
>
>            {
>
>              x <- t.test(ds$Value1[ds$Category == i],
> ds$Value2[ds$Category
> == i])
>
>              y <- c(y, x$p.value)
>
>            }
>
>
>
> This will run, but only if the members of the data frame I am passing as
> an
> argument are in fact named Value1, Value2, and Category.  This is fine for
> now, but in the future I will have to run this function on data where I
> cannot be sure this is the case.  Rather than change the member names by
> hand, I would like to make the function generic to work with any data
> frame.
> How do I do this?  Or is there a better way to do this without the for()
> loop (for example, using apply())?
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>


-- 
Henrique Dallazuanna
Curitiba-Paraná-Brasil
25° 25' 40" S 49° 16' 22" O

	[[alternative HTML version deleted]]

Hi David,

Try this:

# Data set
set.seed(123)
Category=as.factor(rep(1:15,each=10))
Value1 = rnorm(150)
Value2= rnorm(150)
yourdata=data.frame(Category,Value1,Value2)

# Global function
TTEST=function(mydata){

# Internal function
tt=function(x,y) t.test(x,y)$p.value

# p-values
for(i in 1:length(levels(mydata$Category))){
mydatai=mydata[mydata$Category==i,][,-1]
res[i]=tt(mydatai[,1],mydatai[,2])
}

# Result
data.frame(Category=levels(Category),pvalue=res)
}


TTEST(yourdata)
Category     pvalue
1         1 0.88699832
2         2 0.87711367
3         3 0.26075787
4         4 0.30382321
5         5 0.59213871
6         6 0.83755043
7         7 0.47836246
8         8 0.37509850
9         9 0.26132601
10       10 0.29195145
11       11 0.24169206
12       12 0.25594943
13       13 0.34882014
14       14 0.85755554
15       15 0.04556924



HTH,


Jorge



On Fri, May 2, 2008 at 10:19 AM, David Schwab <dvschwab46225@gmail.com>
wrote:
> I am writing a simple R program to execute a t-test repeatedly on data
> contained in a data frame.  My data looks like this:
>
>
>
> Category                      Value1             Value2
>
> 1                                  .5                     .8
>
> 1                                  .3                     .9
>
> . . .                               . . .                   . . .
>
> 2                                  1.4                   1.3
>
> 2                                  1.3                   1.3
>
> . . .                               . . .                   . . .
>
> 15                                .2                     .3
>
> 15                                .5                     .1
>
>
>
>
>
> So in all there are 15 categories, and each category contains two sets of
> observations which I want to compare.  I only want to compare Value1 and
> Value2 within each category, but I need to do it 15 times (once for each
> category), so I wanted to write an R function to make it easier.
>
>
>
> Right now I am using a for() loop to do the comparison.  My loop looks
> like
> this:
>
>
>
> for(i in 1:21)
>
> {
>
>            x <- t.test(Value1[Category == i], Value2[Category == i])
>
>            y <- c(y, x$p.value)
>
> }
>
>
>
> The loop runs and everything is working well.  However, I am not sure how
> to
> translate this code into a function.  In particular, I'm not sure how
to
> write a function that passes a data frame ds (containing Category, Value1,
> and Value2 as members) as an argument, and then accessing these members
> within the body of the function.  I've tried the following:
>
>
>
> repeated_test <- function(data)
>
> {
>
>            for(i in 1:21)
>
>            {
>
>              x <- t.test(ds$Value1[ds$Category == i],
> ds$Value2[ds$Category
> == i])
>
>              y <- c(y, x$p.value)
>
>            }
>
>
>
> This will run, but only if the members of the data frame I am passing as
> an
> argument are in fact named Value1, Value2, and Category.  This is fine for
> now, but in the future I will have to run this function on data where I
> cannot be sure this is the case.  Rather than change the member names by
> hand, I would like to make the function generic to work with any data
> frame.
> How do I do this?  Or is there a better way to do this without the for()
> loop (for example, using apply())?
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]