[datatable-help] Weird behavior with S4 subclasses of data.table after loading RCurl

[Jeffrey Arnold]

And what's even more unusual is that for the line DT2[ , v],  the function
`[.data.table` is never called.  I included
debug(data.table:::`[.data.table`) and nothing happened for DT2[ , v], but
it was activated for DT2[ , 2] and the other evaluations of extract on DT2.
 My hunch is that the since the S4 method is dispatching the method with a
signature that includes the classes for i and j, something is going wrong
there since it is only getting awry when the class of j is a name.  I'm
going to test out that hunch.

On Sat, Sep 29, 2012 at 9:40 PM, Jeffrey Arnold <jeffrey.arnold at gmail.com>wrote:

> While I came across this while working on a package that used S4, the code
> in the OP has those errors when run in the global environment.  In the
> output below, I put data.table:::cedta() before and after loading RCurl,
> and it returns TRUE in both cases. The data.table object DT1 works fine
> before and after loading RCurl; DT2, object of the S4 class inheriting from
> data.table, is acting like a data.table object and not like a data.frame in
> all respects that I've looked at, except when I use an unquoted variable
> name in i. That works fine before loading RCurl, but fails after it. I have
> no idea what RCurl could be changing that would affect this, which is why I
> am stumped.
>
> > library("data.table")
> data.table 1.8.2  For help type: help("data.table")
> > sessionInfo()
> R version 2.15.1 (2012-06-22)
> Platform: x86_64-unknown-linux-gnu (64-bit)
>
> locale:
>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
>  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
>  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
>  [7] LC_PAPER=C                 LC_NAME=C
>  [9] LC_ADDRESS=C               LC_TELEPHONE=C
> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] data.table_1.8.2
> > ## cedta is TRUE
> > data.table:::cedta()
> [1] TRUE
> >
> > setClass("DataTable2", contains="data.table")
> >
> > DT1 <- data.table(x=rep(c("a","b","c"),each=3), y=c(1,3,6), v=1:9)
> > DT2 <- new("DataTable2", data.table(x=rep(c("a","b","c"),each=3),
> y=c(1,3,6), v=1:9))
> >
> > library("RCurl")
> Loading required package: bitops
> > sessionInfo()
> R version 2.15.1 (2012-06-22)
> Platform: x86_64-unknown-linux-gnu (64-bit)
>
> locale:
>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
>  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
>  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
>  [7] LC_PAPER=C                 LC_NAME=C
>  [9] LC_ADDRESS=C               LC_TELEPHONE=C
> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] RCurl_1.95-0     bitops_1.0-4.1   data.table_1.8.2
> > ## cedta is still TRUE
> > data.table:::cedta()
> [1] TRUE
> >
> > ## This still works
> > DT1[,v]
> [1] 1 2 3 4 5 6 7 8 9
> > ## This no longer works
> > DT2[,v]
> Error: object 'v' not found
> >
> > ## DT2 still behaving like data.table and not data.frame in other
> respects
>  > DT2[ , 4]
> [1] 4
> > DT2[ , "v"]  # returns v instead of the column
> [1] "v"
> > DT2[ , "v", with=FALSE] # returns the column
>    v
> 1: 1
> 2: 2
> 3: 3
> 4: 4
> 5: 5
> 6: 6
> 7: 7
> 8: 8
> 9: 9
>
>
>
> On Sat, Sep 29, 2012 at 8:34 PM, Steve Lianoglou <
> mailinglist.honeypot at gmail.com> wrote:
>
>> Hi,
>>
>> On Sat, Sep 29, 2012 at 6:21 PM, Matthew Dowle <mdowle at mdowle.plus.com>
>> wrote:
>> >
>> > Hi,
>> >
>> > Quick answer to start. Assuming your package Imports or Depends on
>> > data.table (in the DESCRIPTON file) then see here for description of
>> > `cedta` and how it works and maybe it needs a fix :
>> >
>> > http://stackoverflow.com/a/10529888/403310
>>
>> Quick note: this doesn't look like it has to do w/ cedta ... I was
>> debugging this w/ the sample provided and cedta() returns TRUE.
>>
>> Somehow ~ line 780 of data.table.R, `xvars` is empty and I guess the
>> column is injected into the SDenv befure the `jval = eval(jsub,SDenv)`
>> call, so its not found.
>>
>> -steve
>>
>> --
>> Steve Lianoglou
>> Graduate Student: Computational Systems Biology
>>  | Memorial Sloan-Kettering Cancer Center
>>  | Weill Medical College of Cornell University
>> Contact Info: http://cbio.mskcc.org/~lianos/contact
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/datatable-help/attachments/20120929/6a3caff6/attachment-0001.html>