Integer variables not handled by some functions

[pchelle]

Sometimes observed data are assumed as integer and not numeric variables by R, which crashes the function toUnit with the error message below. It could be helpful to enforce integer variables into numeric within functions such as toUnit (using as.numeric ?)

ospsuite::toUnit("Time", as.integer(c(1,2,3)), "h")

Error in rClr::clrCall(dimensionTask, "ConvertToUnit", dimension, targetUnit,  : 
  Type:    System.MissingMethodException
Message: Could not find a suitable instance method ConvertToUnit on type OSPSuite.R.Services.DimensionTask for method parameters String, String, Int32[]
Method:  Void ThrowMissingMethod(System.Type, System.String, System.String, System.Type[])
Stack trace:
   at Rclr.ReflectionHelper.ThrowMissingMethod(Type classType, String methodName, String modifier, Type[] types)
   at Rclr.ClrFacade.InternalCallInstanceMethod(Object obj, String methodName, Boolean tryUseConverter, Object[] arguments)

[Yuri05]

ospsuite::toUnit("Time", as.integer(1), "h") works though...

[PavelBal]

Where do you get values that are integer and not numeric from? Just curious what is the use case.

[pchelle]

When reading an observed dataset from a csv file and all the values within a column are integers, R assumes that their class is integer and not numeric.
Since before plotting the data, we need to convert them to a display unit, it creates this crash on some test examples.

[pchelle]

More specifically it is usually the time variable of observed data which includes only integers

[PavelBal]

@pchelle could you provide an example csv file and the code how you read it?

[pchelle]

To read the observed data, RE function is readObservedDataFile but it is based on

read.csv(fileName,
      header = TRUE,
      check.names = FALSE,
      encoding = "UTF-8"
    )

[pchelle]

The example is in the zip
Itraconazole 600mg MD.zip

[msevestre]

@pchelle Fixed. Updating reportingengine with this build