Error while using R.ScriptEval()

[123saga]

Description

I am using R.ScriptEval() to perform clustering analysis using a source R file and assign colors to data points.

From plugin console, I can see that:

• I connected to R plugin
• Successfully executed the R script
• But failed to receive response from plugin

Steps to Reproduce

expression in in Qlik to assign color to data points in a scatter plot

R.ScriptEval('source("/path/POC_Predict.R")', [param1] as param 1, [param2] as param 2, ... [paramn] as param n)

# read params passed by R.scriptEval()
# do something
# last line of the R Script
as.data.frame(response$predict)

Expected behavior

data points are colored as per the cluster index received from R

Actual behavior

Error: I see the following error in plugin console log:

2018-03-22 18:42:31.0128|INFO|SSEtoRserve.RServeEvaluator|Rserve result: 2 rows, hashid (41876889) 2018-03-22 18:42:31.0128|DEBUG|SSEtoRserve.RServeEvaluator|Rserve result column names: value, visible 2018-03-22 18:42:31.0128|DEBUG|SSEtoRserve.RServeEvaluator|Rserve result nrOfCol umns: 2 2018-03-22 18:42:31.0128|WARN|SSEtoRserve.RServeEvaluator|Rserve result, column data type not recognized: RserveCLI2.SexpList

Environment

Plugin is installed on a Windows VM

Operating system where the SSE plugin is deployed

[x] Windows
[ ] Linux

Qlik Sense

[ ] Desktop
[x] Server

Versions

• sse-r-plugin: Qlik Sense February 2018 | v1.1.0
• Qlik Sense: QS Nov 2017.
• Operating system: Windows Server 2012 R2
• [Other relevant versions] I am running R server and SSE plugin on a different VM, the sample k-mean clustering example with iris data is working fine, please let me know if you any other details.

Thanks in advance.

[123saga]

Hi @tobiaslindulf , Could you please help me with the above issue when you get a chance, let me know if you need more information.

[tobiaslindulf]

Hi, thanks for reaching out for help. Interesting problem. I can think of some different things to look into:

1. Do you expect as.data.frame(response$predict) to produce a data frame with 2 columns? If not then try to return just 1 column by modifying the R script.
2. Can you provide a real example of how your call to R.ScriptEval('source("/path/POC_Predict.R")', [param1] as param 1, [param2] as param 2, ... [paramn] as param n) looks like. If you use spaces in the names you may get problems.
3. Can you try to use the latest version of the SSEtoRServe plugin (1.2.0). It is backwards compatible with your sense version. That plugin version handles data frames from R a bit better than before.
4. your script argument of source is probably not returning anything so you need to call something in that local R file as well, like a function with q as input. Also, I think you need to set the working directory to something just in case it is not set in the Rserve session. Like this: if you have a local R file named myEcho.R containing this: myEcho <- function(df) { df }

Then the R script in the expression could look like this: setwd("C:/Users/userX/Documents/R/Myscripts"); source("myEcho.R"); myEcho(q);

I hope this helps.

[123saga]

@tobiaslindulf , thanks for the response, I will try 1, re-check 2, try 3 and for 4: i think I can run my script stand-alone, so this should not be a problem, I will check this again and get back to you, thanks!