sparklyr / sparklyr

R interface for Apache Spark

Home Page:https://spark.rstudio.com/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

"Error in if: argument is of length zero" shows when creating a new connection in workbench with Spark 3.5, 3.4.1 and 3.3.3

astayleraz opened this issue · comments

Description:
I see that when I install spark 3.5, 3.4.1 or 3.3.3 and then try and add a new connection from the connections pane, I see these warnings and error in the console, and in the spark version drop down it does not show those versions Im trying. In the console, I can create the connection with sc <- spark_connect(master = "local")

Screenshot 2023-10-12 at 2 38 42 PM

Test steps:

  1. Open an RStudio session in workbench and load the sparklyr. library(sparklyr)
  2. In the workbench terminal download spark 3.5.0 (happens with 3.4.1 and 3.3.3 as well) wget -O spark-3.5.0-bin-hadoop3.tgz https://rstudio-workbench-qa.s3.us-west-2.amazonaws.com/spark-3.5.0-bin-hadoop3.tgz
  3. In the console, install that. spark_install_tar("spark-3.5.0-bin-hadoop3.tgz")
  4. From the connections pane, click New Connection and select Spark.

Expected:
I would not expect any errors, and I would expect that version of spark to show as an option in the dropdown.

Actual:
I see the following warnings and error. This is from when I was testing 3.3.3.

Listening on http://127.0.0.1:4819
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'
Warning: Error in if: argument is of length zero
  68: paste
  67: generateCode [/opt/R/4.2.2/lib/R/library/sparklyr/rstudio/shinycon/app.R#270]
  66: <reactive:codeReactive> [/opt/R/4.2.2/lib/R/library/sparklyr/rstudio/shinycon/app.R#320]
  50: codeReactive
  46: observe [/opt/R/4.2.2/lib/R/library/sparklyr/rstudio/shinycon/app.R#342]
  45: <observer>
   2: shiny::runApp
   1: shiny::runGadget
Warning in sprintf(versions$pattern, version$spark, version$hadoop) :
  2 arguments not used by format 'spark-3.3.3-bin-hadoop3'

Note: Jonathan mentioned the error coming from here

componentName <- sub("\\.tgz", "", sprintf(versions$pattern, version$spark, version$hadoop))

Hi @astayleraz , thank you I'll mark this one as a bug to work on