Calling Variables in a Formula in r - r

Question

I want to generalize my function in. I have defined form = y~x, and want to call y so that I use it within the function. y should be user (dynamic) defined and that is why I need a way to call it from the form. This is part of the code I tried and newdata is calculated within the function.
form = y ~ x
newdata = y
trial = function(form, x){
y = newdata
reg = lm(form, data = data.frame(x, newdata))
reg
}

Answer 1

If you want your function to work you could try :
form = y ~ x
newdata = 3*(1:100)+2
trial = function(f=form, x){
y = newdata
reg = lm(f, data = data.frame(x, y=newdata))
reg
}
trial(x=1:100) # or trial(form,1:100)
Call:
lm(formula = f, data = data.frame(x, y = newdata))
Coefficients:
(Intercept) x
2 3 # as expected
The think is, in your previous function, the form was an argument and R did not understand that it was the form you defined. Using f with a default value of form solves this issue.
Was that what you wanted ?
Note that if you want to call the function and define y in the call you could do :
form = y ~ x
trial = function(f=form, x, y){
reg = lm(f, data = data.frame(x, y))
reg
}
trial(x=1:100,y=3*(1:100)+2)
Call:
lm(formula = f, data = data.frame(x, y))
Coefficients:
(Intercept) x
2 3
From what I understood from your comments this might be closer to what you expected :
form = y ~ x
trial = function(f=form, x, y){
y = 5*y+2
reg = lm(f, data = data.frame(x, y))
reg
}
trial(x=1:100,y=3*(1:100))
Call:
lm(formula = f, data = data.frame(x, y))
Coefficients:
(Intercept) x
2 15
You call the function specifying x and giving a "first value" for y. Then, in the function, y is transformed (quite an easy transformation here) and then the regression is done.
Note that when you call the function, y is not the variable y but the argument in the function. If you have a variable z equal to 3*(1:100) you can do trial(x=1:100,y=z)

Answer 2

I tried this out and it turned out to be what I needed
form = y ~ x
trial = function(form, x){
.
.
.
newdata = something
.
.
new.form <- as.formula(call("~", form[[2]], form[[3]]))
assign(deparse(form[[2]]), newdata)
reg = lm(new.form, data = data.frame(x, newdata))
reg
}

Answer 3

Something like this?
ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
group <- gl(2, 10, 20, labels = c("Ctl","Trt"))
weight <- c(ctl, trt)
lm(weight ~ group)
form <- as.formula(y ~ x)
trial = function(form,y, x){
reg = lm(form, data = data.frame(x=x, y=y))
reg
}
trial(form,weight,group)