R help - Feb 2011 - accessing variables inside a function, inside a loop

dear list!
I KNOW this has been answered a million times before. But, as some might 
remember from their "freelance" times as a statistic consultant, 
sometimes you're blinded by the facts. I KNOW I have seen this problem 
solved, but due to the mental blockade i have been suffering from the 
last 3 hours, I can neither find the answer on the mailing list nor on 
google. I wrote a function to plot one continuous variable against an 
abundance of fish, the data is called water. however, my data covers 9 
parameters, and since I am lazy as hell, I decided to write a for loop 
to pass all the variables to the function. as usual, the variables and 
the data frame cannot be found inside the function.

water<-read.csv("Henni/water-abundance2.csv",sep="\t",header=T)
source("functions.R")
attach(water)

names(water):
  [1] "site"     "sample"   "temp"    
"pH"       "DO"       "BOD"
  [7] "COD"      "no3"      "no2"     
"po4"      "N_male"   "N_female"

vars<-names(water)[3:10]

par(mfrow=c(4,2))
for (i in vars)
{
     comp.plot(i,N_male,DATA=water)
}

comp.plot<-function(parameter,y,DATA){ #this is actually in functions.R
     x<-DATA[[parameter]]
     y<-DATA[[y]]
     plot(y~parameter)
}

and of course, this doesn't work. I am really frustrated by now, since I 
have gathered all I could recollect from my past 2 years of R experience 
but still, I was not able to find a solution. I'm really dizzy right 
now, but since I will not be able to go to sleep unless I have found the 
answer to this problem, I urge someone to shed some light on my trivial 
problem! I will promise not to post this question ever again, but for 
now, I'm just stuck!

Thanks in advance and humble apologies to all that feel offended by my post!

A quick look says you are passing N_Male in which is a vector of
values, but you are trying to use it as if it were a character value
in:

y <- DATA[[y]]

try calling it with

comp.plot(i,"N_male",DATA=water)

Also get out of the habit of using 'attach'; it can lead to other
issues.

On Sun, Feb 27, 2011 at 3:48 PM, Sacha Viquerat <tweedie-d at web.de>
wrote:
> dear list!
> I KNOW this has been answered a million times before. But, as some might
> remember from their "freelance" times as a statistic consultant,
sometimes
> you're blinded by the facts. I KNOW I have seen this problem solved,
but due
> to the mental blockade i have been suffering from the last 3 hours, I can
> neither find the answer on the mailing list nor on google. I wrote a
> function to plot one continuous variable against an abundance of fish, the
> data is called water. however, my data covers 9 parameters, and since I am
> lazy as hell, I decided to write a for loop to pass all the variables to
the
> function. as usual, the variables and the data frame cannot be found inside
> the function.
>
>
water<-read.csv("Henni/water-abundance2.csv",sep="\t",header=T)
> source("functions.R")
> attach(water)
>
> names(water):
> ?[1] "site" ? ? "sample" ? "temp" ? ?
"pH" ? ? ? "DO" ? ? ? "BOD"
> ?[7] "COD" ? ? ?"no3" ? ? ?"no2" ? ?
?"po4" ? ? ?"N_male" ? "N_female"
>
> vars<-names(water)[3:10]
>
> par(mfrow=c(4,2))
> for (i in vars)
> {
> ? ?comp.plot(i,N_male,DATA=water)
> }
>
> comp.plot<-function(parameter,y,DATA){ #this is actually in functions.R
> ? ?x<-DATA[[parameter]]
> ? ?y<-DATA[[y]]
> ? ?plot(y~parameter)
> }
>
> and of course, this doesn't work. I am really frustrated by now, since
I
> have gathered all I could recollect from my past 2 years of R experience
but
> still, I was not able to find a solution. I'm really dizzy right now,
but
> since I will not be able to go to sleep unless I have found the answer to
> this problem, I urge someone to shed some light on my trivial problem! I
> will promise not to post this question ever again, but for now, I'm
just
> stuck!
>
> Thanks in advance and humble apologies to all that feel offended by my
post!
>
> ______________________________________________
> R-help at 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.
>


-- 
Jim Holtman
Data Munger Guru

What is the problem that you are trying to solve?

Dear Sacha,

That is the most piteous soliloquy I have read in some time.  Here
then a response to soothe your dizzied mind and wrest your rest from
the ornery sprites that so threaten to plague it now.

## Some sample data
water <- as.data.frame(matrix(rnorm(120), ncol = 12))
## Adding your names
names(water) <- c("site", "sample", "temp",
"pH", "DO", "BOD",
  "COD", "no3", "no2", "po4",
"N_male", "N_female")

par(mfrow = c(4, 2))

## Simplest
lapply(water[, 3:10], FUN = plot, y = water[, "N_male"])

## Slightly fancier
plotter <- function(x) {
  form <- eval(substitute(N_male ~ p, list(p = as.name(x))))
  plot(form, data = water)
}

##
lapply(names(water[, 3:10]), plotter)

## Or if you wanted to start getting "really" fancy
## you could wrap it all up

jwrap <- function(y, x, dat, mfrow) {
  old.par <- par(no.readonly = TRUE)
  on.exit(old.par)
  par(ask = TRUE, mfrow = mfrow)

  jplot <- function(x, y) {
    form <- eval(substitute(y ~ p, list(y = as.name(y), p = as.name(x))))
    plot(form, data = water)
  }

  lapply(X = names(dat[, x]), FUN = jplot, y = y)
}

## jwrap() in action
jwrap("N_male", x = 3:10, dat = water, mfrow = c(4, 2))
## and showing what happens if you change mfrow
jwrap("N_male", x = 3:10, dat = water, mfrow = c(2, 2))

In all good cheer,

Joshua

On Sun, Feb 27, 2011 at 12:48 PM, Sacha Viquerat <tweedie-d at web.de>
wrote:
> dear list!
> I KNOW this has been answered a million times before. But, as some might
> remember from their "freelance" times as a statistic consultant,
sometimes
> you're blinded by the facts. I KNOW I have seen this problem solved,
but due
> to the mental blockade i have been suffering from the last 3 hours, I can
> neither find the answer on the mailing list nor on google. I wrote a
> function to plot one continuous variable against an abundance of fish, the
> data is called water. however, my data covers 9 parameters, and since I am
> lazy as hell, I decided to write a for loop to pass all the variables to
the
> function. as usual, the variables and the data frame cannot be found inside
> the function.
>
>
water<-read.csv("Henni/water-abundance2.csv",sep="\t",header=T)
> source("functions.R")
> attach(water)
>
> names(water):
> ?[1] "site" ? ? "sample" ? "temp" ? ?
"pH" ? ? ? "DO" ? ? ? "BOD"
> ?[7] "COD" ? ? ?"no3" ? ? ?"no2" ? ?
?"po4" ? ? ?"N_male" ? "N_female"
>
> vars<-names(water)[3:10]
>
> par(mfrow=c(4,2))
> for (i in vars)
> {
> ? ?comp.plot(i,N_male,DATA=water)
> }
>
> comp.plot<-function(parameter,y,DATA){ #this is actually in functions.R
> ? ?x<-DATA[[parameter]]
> ? ?y<-DATA[[y]]
> ? ?plot(y~parameter)
> }
>
> and of course, this doesn't work. I am really frustrated by now, since
I
> have gathered all I could recollect from my past 2 years of R experience
but
> still, I was not able to find a solution. I'm really dizzy right now,
but
> since I will not be able to go to sleep unless I have found the answer to
> this problem, I urge someone to shed some light on my trivial problem! I
> will promise not to post this question ever again, but for now, I'm
just
> stuck!
>
> Thanks in advance and humble apologies to all that feel offended by my
post!
>
> ______________________________________________
> R-help at 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.
>


-- 
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/

Joshua

Great solution.  Taking off on your code, the following works but does not
display the names of the variables in the formula.  Any suggestions about how to
modify the function so that it displays the correct formula (e.g.,
"glm(formula = y1 ~ x1 * x2, data = dat)" instead of "glm(formula
= frm, data = dat)")?

R> x = runif(2000)
R> y = runif(2000)
R> z = runif(2000)
R> 
R> y1 = y * x
R> y2 = y * sqrt(x)
R> 
R> x1 = y1 / y2 + z
R> x2 = y2 / y1 * z + z
R> 
R> dat = data.frame(y1,y2,x1,x2)
R>
R> xReg = function(y) {
+ 
+ 	frm = eval(substitute(p ~ x1 * x2, list(p = as.name(y))))
+ 	mod = glm(frm, data=dat)
+ 	}
R> 
R> lapply(names(dat[,1:2]), xReg)
[[1]]

Call:  glm(formula = frm, data = dat)

Coefficients:
(Intercept)           x1           x2        x1:x2  
-0.1882452    0.4932059    0.0667401   -0.1310084  

Degrees of Freedom: 1999 Total (i.e. Null);  1996 Residual
Null Deviance:	    99.15032 
Residual Deviance: 67.71775 	AIC: -1085.354 

[[2]]

Call:  glm(formula = frm, data = dat)

Coefficients:
(Intercept)            x1            x2         x1:x2  
-0.005464627   0.386937367   0.037363416  -0.094136334  

Degrees of Freedom: 1999 Total (i.e. Null);  1996 Residual
Null Deviance:	    112.7078 
Residual Deviance: 90.24796 	AIC: -510.9287 

---

Thanks,

Alan