[datatable-help] packages data.table and treemap, tabplot and SGP

[Matthew Dowle]

> On 16/07/2012 09:06, Matthew Dowle wrote:
>>
>> Uwe, CRAN maintainers and Martijn,
>> cc datatable-help
>>
>> Sincere apologies. I had checked dependent packages but not before some
>> last minute changes. I should know better to recheck.
>>
>> In addition to treemap and tabplot now failing, it's failing on Mac too.
>> Likely due to new tests with non-ascii characters. I had anticipated
>> encoding issues and coded for that, though. I tested on Linux,
>> winbuilder
>> for Windows, and it passes checks on Mac on R-Forge. So will need to
>> investigate why encoding doesn't pass Mac on CRAN.

Brian Ripley wrote :

> It's not Mac, it is that the CRAN Mac checks are done in the C locale
> for some reason best known to Simon Urbanek.  So you need to check the
> locale before running the tests.

Ok, thank you. Will do.

>
> Also, please check SGP, which now gives
>
> * checking R code for possible problems ... NOTE
> abcSGP: no visible global function definition for ‘J’
> ... many more of the same.

That was the main thing I checked dependent packages for. I am sure SGP
passed when I ran it. I really don't know what happened here.  Honestly I
ran R CMD check on SGP and it passed fine. I wonder if J() was being
defined separately in my environment, somehow. R CMD check runs under
--vanilla, so that can't be it.  Unless ... when I ran the dependent
package checks maybe I hadn't actually R CMD INSTALL'd the very latest
version of data.table. Or had reverted to a previous version to test
something and had forgotten to install again.  Oh no ... I bet that was
it.

I think I can still get a fix for this too, tonight. Will have to keep J
defined and allow longer to deprecate it. There is a detailed NEWS item
about it, and advance warning to datatable-help, but I mistakenly believed
dependent packages didn't use it outside DT[].

For clarity ... no dependent packages should need to change. It's all
data.table's fault. Will submit 1.8.3 asap.

Matthew


>>
>> I'll endeavour to subsmit v1.8.3 tonight or tomorrow.
>>
>> Apologies again,
>> Matthew
>>
>>> Dear package maintainers of data.table and treemap, tabplot,
>>>
>>> probably due to yesterday's data.table update on CRAN, we see treemap
>>> and tabplot failing now, this will shortly be reported on
>>>
>>> http://cran.r-project.org/web/checks/check_results_treemap.html
>>> http://cran.r-project.org/web/checks/check_results_tabplot.html
>>>
>>> Can you please negotiate further actions and fix the problem. Please
>>> upload fixed version(s) of the package(s) with increased version
>>> number(s) to CRAN in the usual way and report to cran at R-project.org.
>>>
>>> In principle, according to CRAN policies, we have to revert the
>>> data.table update, since it is the maintainer's responsibility to check
>>> an update breaks other packages.
>>> We will revert if we do not get a response to this message within 24
>>> hours from now.
>>>
>>>
>>> For now, the relevant part of the logfiles are, for treemap:
>>>
>>> * checking examples ...
>>> ** running examples for arch 'i386' ... ERROR
>>> Running examples in 'treemap-Ex.R' failed
>>> The error most likely occurred in:
>>>
>>>   > ### Name: tmPlot
>>>   > ### Title: Create treemap
>>>   > ### Aliases: tmPlot
>>>   >
>>>   > ### ** Examples
>>>   >
>>>   > #########################################
>>>   > ### quick example
>>>   > ###
>>>   > ### data: Gross national income data
>>>   > #########################################
>>>   >
>>>   > # load Gross national income data
>>>   > data(GNI2010)
>>>   >
>>>   > # create treemap
>>>   > tmPlot(GNI2010,
>>> +        index=c("continent", "iso3"),
>>> +        vSize="population",
>>> +        vColor="GNI",
>>> +        type="value")
>>>   >
>>>   >
>>>   >
>>>   > #########################################
>>>   > ### extended examples
>>>   > ###
>>>   > ### data: fictive structural business statistics data
>>>   > #########################################
>>>   >
>>>   > ### load fictive structural business statistics data
>>>   > data(sbsData)
>>>   > sbsData$employees.growth <- sbsData$employees09 -
>>> sbsData$employees08
>>>   >
>>>   >
>>>   > #########################################
>>>   > ### types
>>>   > #########################################
>>>   >
>>>   > # value treemap: the color variable is directly mapped to the
>>> colors
>>>   > tmPlot(sbsData,
>>> +        index=c("section", "subsection"),
>>> +        vSize="employees09",
>>> +        vColor="employees.growth",
>>> +        type="value")
>>> Error in rbindlist(allargs) :
>>>     Column 1 of item 2 is type 'integer', inconsistent with column 1 of
>>> item 1's type ('double')
>>> Calls: tmPlot ... baseTreemap -> rbind -> rbind -> .rbind.data.table ->
>>> rbindlist
>>> Execution halted
>>>
>>>
>>>
>>>
>>> for tabplot:
>>>
>>> ** running examples for arch 'i386' ... ERROR
>>> Running examples in 'tabplot-Ex.R' failed
>>> The error most likely occurred in:
>>>
>>>   > ### Name: tableChange
>>>   > ### Title: Change a tabplot-object
>>>   > ### Aliases: tableChange
>>>   >
>>>   > ### ** Examples
>>>   >
>>>   > # load diamonds dataset from ggplot2
>>>   > require(ggplot2)
>>> Loading required package: ggplot2
>>>   > data(diamonds)
>>>   >
>>>   > # assign tableplot as tabplot object
>>>   > tab <- tableplot(diamonds, plot=FALSE)
>>> Error in tabulate(get(col), nbins = nlev) :
>>>     'bin' must be numeric or a factor
>>> Calls: tableplot ... getFreqTable_DT -> matrix -> [ -> [.data.table ->
>>> tabulate
>>> Execution halted
>>>
>>>
>>>
>>> Best wishes,
>>> Uwe Ligges
>>>
>>
>>
>
>
> --
> Brian D. Ripley,                  ripley at stats.ox.ac.uk
> Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
> University of Oxford,             Tel:  +44 1865 272861 (self)
> 1 South Parks Road,                     +44 1865 272866 (PA)
> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>
>
>