topic Re: R interface in JMP in Discussions

R interface in JMP

CheeseProgram — Sat, 10 Jun 2023 23:48:08 GMT

Hello,

I wrote a code in R that I want to run into JMP. Here is the code :

R init();

//send data to R
R Send(dt);
R submit("
library(spatstat)
#creation fenetre
fenn<-owin(xrange=c(40,760), yrange=c(40,160), poly=NULL, mask=NULL, units=NULL)
ppp<-ppp(x=dt$XM,y=dt$YM,window=fenn)
print(quadratcount(ppp,nx=6,ny=3)

Q<-quadrat.test.ppp(ppp,nx=6,ny=3)

a<-cdf.test(ppp,'x')

D=0
if(a[[2]] & Q[[3]] < 0.1 ){
  D=0}else {D=1}
Distribution=as.data.frame(rep(c(D),nrow(dt)))

class(Distribution)
print(a)

");
R term();

Then when I run it I got two issues:

Firstly I have an error from these lines :

fenn<-owin(xrange=c(40,760), yrange=c(40,160), poly=NULL, mask=NULL, units=NULL)
ppp<-ppp(x=dt$XM,y=dt$YM,window=fenn)

which says :

Error: Function boundingbox called with 2 unrecognised arguments
In addition: Warning message:
2 points were rejected as lying outside the specified window

In fact when I change 40 into 0 it works because all my points are in my window, but when I run the code on R the fact that all my points are not in the window is not an issue. And if I run the previous code in R it works. So I don't understand why I got a warning message in JMP but not in R.

The second issue is when I want to export my data, JMP don't want to extract my data frame and I don't understand why. I use this code :

ResidJMP = R Get(Distribution);
ResidJMP << New Data View;
R term();

In fact I want to return a data table in JMP and after I want to use it as the main data table.

Re: R interface in JMP

flo_kussener — Mon, 09 May 2022 15:29:16 GMT

Hello,

For a basic example,

R Init();   // Initiate R connection

dt = Open(“$SAMPLE_DATA/Big Class.jmp”,invisible);
R Send(dt);

the call of column should be done on the following way:

R Submit("
print(dt['weight'])
	//b<-dt['weight']/dt['height']
");

instead of

R Submit("
print(dt$weight)
	//b<-dt['weight']/dt['height']
");

I think it should change the output of your first problem.

Re: R interface in JMP

CheeseProgram — Mon, 09 May 2022 15:52:06 GMT

When I apply this to my data it says that the columns I used aren't numeric, then when I use as.numeric() function it says : 'list' object cannot be coerced to type 'double'. This problem doesn't appeared with the $ call. There are so many issues here unfortunately

Re: R interface in JMP

Mark_Bailey — Mon, 09 May 2022 16:26:16 GMT

You can use a JMP data table as a data frame in R. The R Send() function automatically converts one data structure into the other. Here is a simple example:

Names Default to Here( 1 );

// make sure JMP can find the desired installed version of R
Set Environment Variable( "R_HOME", "C:\Program Files\R\R-4.1.3" );

// use regression example from JMP Sample Data Folder
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );

// use standard linear model function with data frame
R Init( );
R Send( dt );
R Submit(
"fit <- lm( dt$weight~dt$height )
print( fit )"
);
R Term();

// examine results in Log

So I suspect that there is a problem with the R code (function arguments), or the nature of the variables in the data frame. After you make sure that your function call is correct, try printing the data table from R to see if everything transferred as expected.

Re: R interface in JMP

CheeseProgram — Mon, 09 May 2022 16:44:08 GMT

When I use this code on R

library(spatstat)
data<-read.("Results.csv")
#creation fenetre
fenn<-owin(xrange=c(40,760), yrange=c(40,160), poly=NULL, mask=NULL, units=NULL)
#creation point pattern
ppp<-ppp(x=data$XM,y=data$YM,window=fenn)
plot(ppp)
summary(ppp)

#test quadrat
quadratcount(ppp,nx=6,ny=3)
Q<-quadrat.test.ppp(ppp,nx=6,ny=3)
paste("Ma pvalue du quadrat test ", format(Q[[3]], scientific=TRUE, digits=3))

#test ks
a<-kstest(ppp,'x')

paste("Ma pvalue du KS test ", format(a[[2]], scientific=TRUE, digits=3))

D=character()
if(a[[2]] & Q[[3]] < 0.1 ){
  D="Distribution Non-Uniforme"}else {D="Distribution Uniforme"}
Distribution<-rep(c(D),length(XM))
dt<-cbind(dt,Distribution)

It works good. In fact it works good in JMP if I use this window :

fenn<-owin(xrange=c(0,760), yrange=c(0,160), poly=NULL, mask=NULL, units=NULL)

Because all my (x,y) points are in the window, the problem comes out when I put the window I want. But in R there is no problem of window.

But if there is a problem with the variables it shouldn't work with an other window and this is what I don't understand.

Maybe there is something I don't understand in your comment but I don't find what could be the error. When I print it on JMP from R I have the good data frame. I'm lost for sure

Re: R interface in JMP

CheeseProgram — Mon, 09 May 2022 17:11:50 GMT

I try to use R Submit File() function so I can use the exact same code. But I have the same error on the jmp code that I don't have in R. This is the error that I need resolve. I try so much things and also don't understand why it appears.

Error: Function boundingbox called with 2 unrecognised arguments

Re: R interface in JMP

Mark_Bailey — Mon, 09 May 2022 17:12:38 GMT

Did you try my simple example? Does it work? What do you see in the JMP Log?

Did you try sending a JMP data table to R and printing it in R? Does that data frame match the data table?

Re: R interface in JMP

CheeseProgram — Mon, 09 May 2022 17:29:49 GMT

Your example works I had this in the log :

For your second question I'm not sure I'm understanding it well. But I try to send my JMP data table to R with send() command and then print it in the log. For your example it looks like this :

In my example i got this in the log :

and my table look like this :

Re: R interface in JMP

CheeseProgram — Tue, 10 May 2022 10:10:46 GMT

Now my main problem is to export my data frame from R. I have the same error with your example than with my data.

Erreur : Error in if (is.na(datam)) { : the condition has length > 1
Unexpected errors occurred while attempting to transfer the data.

I check that my R get() is on a data.frame. Here is my code :

Treat=Open( "C:\Users\GROUSD01\Desktop\CréationcodeMPPT\Results 4.jmp" );


R init();

//send data to R
R Send(Treat);
R submit("library(spatstat)
round(Treat$XM)
round(Treat$YM)
XM<-Treat$XM
YM<-Treat$YM
test<-as.data.frame(cbind(XM,YM))
test2<-data.frame()
for (i in 1:nrow(test)){
  if(test[i,2]>39){
    test2<-rbind(test2,test[i,1:2])}}


fenn<-owin(xrange=c(40,760), yrange=c(40,160), poly=NULL, mask=NULL, units=NULL)
#creation point pattern
ppp<-ppp(x=test2$XM,y=test2$YM,window=fenn)
print(ppp)
plot(quadratcount(ppp,nx=6,ny=3))
Q<-quadrat.test.ppp(ppp,nx=6,ny=3)
print(Q)

a<-cdf.test(ppp,'x')

paste('Ma pvalue du KS test ', format(a[[2]], digits=3))

D=character()
if(a[[2]] & Q[[3]] < 0.1 ){
  D='Distribution Non-Uniforme'}else {D='Distribution Uniforme'}
Distribution<-(rep(c(D),length(Treat$XM)))
Treat<-cbind(Treat,Distribution)

");
R Get(Treat)<<New Data View
R term();

But I tried a lot of different things for the same results

Re: R interface in JMP

Mark_Bailey — Tue, 10 May 2022 16:18:25 GMT

I suggest that you contact JMP Technical Support (support@jmp.com) for help with this problem.

Re: R interface in JMP

CurseOfCapybara — Mon, 26 Sep 2022 18:30:58 GMT

Was there a solution to the error?

Error in if (is.na(datam)) { : the condition has length > 1

I have the same issue.