Question

Rcaller Output having issues with a list of output attributes

  • 28 April 2022
  • 9 replies
  • 31 views
R
Badge

Hi,

I want to run some Rscript and so installed an R-interpreter and installed some Libraries. As a test case I downloaded the workspace from a rather old article "RCaller: Ins and outs of using R in FME". The workspace is from 2020 though so i ran it on both 2019 and 2021 versions but it returned the same error in both cases. Actually the exact same error as mentioned in the article itself: "Failed with exit code 1 when executing R script. Output was: Loading required package: gsubfn.......". Now the article mentions to check spelling but as the workspace is an exact copy from safe I expect that to be tested and ok and furthermore that is just not the issue as the line "fmeOutput<-data.frame(meanEst,meanAct, r2, m, k)" in the Rcaller works just fine if I just pick one of the columns for output, irrespective which one. e.g. "fmeOutput<-data.frame(r2)" But the moment I comma separate a few data frame columns for output I get the error.

Any help with this appreciated.

All the best, Robbert

 

PS. also ran the code in RStudio where the code as-in the RCaller transformer works fine

9 replies

M
Badge +2

@rokop​ I ran the workspace template rcallerlinearregression.fmwt using FME 2019.2 and it seems to work OK.

imagePerhaps check that FME is pointing to the expected R Interpreter .

imageThe example doesn't require any special packages to be loaded. rcallerlinearregressionwithgroups.fmwt also seems to work fine in FME 2019.2

R
Badge

@rokop​ I ran the workspace template rcallerlinearregression.fmwt using FME 2019.2 and it seems to work OK.

imagePerhaps check that FME is pointing to the expected R Interpreter .

imageThe example doesn't require any special packages to be loaded. rcallerlinearregressionwithgroups.fmwt also seems to work fine in FME 2019.2

Hi,

I must have set the interpreter right as it is all working fine outputting 1 attribute to fmeOutput. Also tried both the 32 and 64 bit versions of Rscript. The error i get:

image" data-fileid="0694Q00000HaWCcQAN 

If take off the encircled attribute meanEst it works and the line where it goes wrong then reads instead:

RCaller (InlineQueryFactory): Processing R script

RCaller (InlineQueryFactory): Generating output 'Output' using query 'SELECT * FROM "Output"'

So it seems to fail on the sql when more than 1 attribute is needed.

Cheers,

Robbert

M
Badge +2

Hi,

I must have set the interpreter right as it is all working fine outputting 1 attribute to fmeOutput. Also tried both the 32 and 64 bit versions of Rscript. The error i get:

image" data-fileid="0694Q00000HaWCcQAN 

If take off the encircled attribute meanEst it works and the line where it goes wrong then reads instead:

RCaller (InlineQueryFactory): Processing R script

RCaller (InlineQueryFactory): Generating output 'Output' using query 'SELECT * FROM "Output"'

So it seems to fail on the sql when more than 1 attribute is needed.

Cheers,

Robbert

@rokop​ What OS are you running? Also what version of R do you have installed?

 

Try using Log Debug under Tools - FME Options - Translation tab - Log Message Filter and attach that log to this conversation.

It is odd that changing the order and number of attributes causes the crash. I don't see that here.

R
Badge

Hi Mark,

I'm on Windows 10 Enterprise. Running R-4.2.0 on fme 2019.2.

Please find attached the run log

 

Cheers,

Robbert

M
Badge +2

Hi Mark,

I'm on Windows 10 Enterprise. Running R-4.2.0 on fme 2019.2.

Please find attached the run log

 

Cheers,

Robbert

@rokop​ Just checking - can you confirm you have installed the sqldf package ?

M
Badge +2

@rokop​ Just checking - can you confirm you have installed the sqldf package ?

Perhaps also check that you do not have two versions of R installed

 

J

Hi, I have the same problem. The transformer only works when the output data frame fmeOutput only contains a single value. Otherwise it fails with the error 

 

Error in if (fmeOutput != "") dbWriteTable(dbcon, "Output", fmeOutput, : 

 the condition has length > 1

 

This screenshot shows it with the only line of code in the RCaller set to:

fmeOutput<-data.frame(numbers=c(1,2,3))

I wonder if an update to one of the packages used by sqldf has caused this to fail. I have the latest version of sqldf installed from Cran today along with dependencies RSQLite, proto, etc.

fme_rcaller

DanAtSafe
Rank
Userlevel 1
Badge +11

Hi, I have the same problem. The transformer only works when the output data frame fmeOutput only contains a single value. Otherwise it fails with the error 

 

Error in if (fmeOutput != "") dbWriteTable(dbcon, "Output", fmeOutput, : 

 the condition has length > 1

 

This screenshot shows it with the only line of code in the RCaller set to:

fmeOutput<-data.frame(numbers=c(1,2,3))

I wonder if an update to one of the packages used by sqldf has caused this to fail. I have the latest version of sqldf installed from Cran today along with dependencies RSQLite, proto, etc.

fme_rcaller

Hi @jo-gay Unfortunately the RCaller doesn't work with R 4.2 - FYI the issue is FMEENGINE-73477

M
Badge +2

@jo-gay​ @rokop​ we think we might have addressed this issue in FME 2022.2 - this will be available as a beta download build 22705 or higher. If someone would like to test it when the beta becomes available.

Reply


Community Stats

28,706
Posts
109,463
Replies
36,549
Members
Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.