fExtremes on the second-last day of the month

Yohan Chalabi chalabi at phys.ethz.ch
Fri Nov 30 10:37:13 CET 2012


Dear Brian,

Thank you for your detailed report.

I have now submitted a new version of the package.

Kind regards,
Yohan

Prof Brian Ripley writes:

> Yohan,
>
> fExtremes fails on the second-last day of the month, and as today is
> such a day and we've worked out the pattern, please take a look today:
>
> * checking tests ...
>    Running ‘doRUnit.R’ [40s/40s]
>   ERROR
> Running the tests in ‘tests/doRUnit.R’ failed.
> Last 13 lines of output:
>    fExtremes unit testing - 47 test functions, 2 errors, 0 failures
>    ERROR in test.gevFitByBlocks: Error in as.matrix(applySeries(x, from,
> to, FUN = which.max)) :
>      error in evaluating the argument 'x' in selecting a method for
> function 'as.matrix': Error: NCOL(rowBind) == NCOL(x) is not TRUE
>
>    ERROR in test.timeSeriesBlocks: Error in as.matrix(applySeries(x,
> from, to, FUN = which.max)) :
>      error in evaluating the argument 'x' in selecting a method for
> function 'as.matrix': Error: NCOL(rowBind) == NCOL(x) is not TRUE
>
>    Error in eval(expr, envir, enclos) :
>
>    unit testing failed (#test failures: 0, R errors: 2)
>
>    Calls: source -> withVisible -> eval -> eval
>
> The problem is to do with missing the last day of the month as the
> period finishes at midnight on the last day (and not at the end of the
> last day), and hence the records are empty.
>
> There are several other issues, not least the maintainer and
>
> Found the following significant note:
>    running .First.lib() for package ‘fExtremes’ as .onLoad/.onAttach
> were not found
>
>
> Brian



More information about the Rmetrics-core mailing list