-
Notifications
You must be signed in to change notification settings - Fork 0
Troubleshooting
Michael McCarthy edited this page Nov 5, 2021
·
5 revisions
The command line will show a "Model Failed" error, along with any error messages passed from other scripts, if an R or Python error is caught. Reasons for these errors may include:
- An R script encountered an error. This is the most common issue, likely caused by a missing input or intermediary file. The error message from R will pass through to the command line window.
- An invalid model component is called in the Batch file. If the batch file is modified, ensure that the model runner and VISUM runner tokens used in the batch file are valid.
- VISUM is not licensed or there is not a compatible license available. In this situation, a VISUM window may open but the software does not allow saving version files. To resolve, make sure the selected license(s) is appropriate for the CALM network size and available to use.
| Issue | Resolution Steps |
|---|---|
| Model does not run | Check the location of the dependencies Check the batch file for valid tokens Check VISUM license Look for any error messages |
| There is an R error | Check the command line window for the error details. An input file may be missing or in the wrong format. Additionally, check that the University model has produced outputs. |
| The University model fails | Check that ZMX files are written to unimodel/emmemat. If there are no files here, check the Settings file for the path to 7-Zip.Check the error log ( /unimodel/rpt/universityModel.log) for detailed errors. If there are errors with a particular model, check the UEC and any input files and skims for that model. |
| VISUM throws an error or behaves unexpectedly | Check for unnecessary matrices or procedure sequence items in the input version file Check VISUM license Check for skim and demand matrices in outputs/matrices
|
| The model run ends earlier than expected | Check max. iterations in the batch file |
| The reporting component fails | Check that referenceRun has been set |
These R scripts should only be run using the version of R software in the dependencies package, which is currently 3.4.1. Running in R versions 3.5 or later may cause new format RData files to be written that are not backwards-compatible. This may cause issues when trying to run the model later via batch script with R version 3.4.1. Output RData files can still be read in later versions of R.