[Rd] Better 'undefined columns' error for data.frame

[Duncan Murdoch]

Hi.  That change is much more ambitious than I would have guessed.  R 
Core might prefer it to arrive via bugzilla.

A couple of small comments:

1.  Your last test, "dfa[,cols]" produces a really long message:

 > dfa[,cols]
Error in `[.data.frame`(dfa, , cols) :
   undefined columns selected:   4,   5,   6,   7,   8,   9,  10,  11, 
12,  13,  14,  15,  16,  17,  18,  19,  20,  21,  22,  23,  24,  25, 
26,  27,  28,  29,  30,  31,  32,  33,  34,  35,  36,  37,  38,  39, 
40,  41,  42,  43,  44,  45,  46,  47,  48,  49,  50,  51,  52,  53, 
54,  55,  56,  57,  58,  59,  60,  61,  62,  63,  64,  65,  66,  67, 
68,  69,  70,  71,  72,  73,  74,  75,  76,  77,  78,  79,  80,  81, 
82,  83,  84,  85,  86,  87,  88,  89,  90,  91,  92,  93,  94,  95, 
96,  97,  98,  99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 
110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 
124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 
138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 
152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 
166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 
180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 
194, 195, 196

but it is still incomplete, because cols is 1:1e6.  It might be 
preferable to use ellipsis when the number of columns is too large, e.g.

 > dfa[,cols]
Error in `[.data.frame`(dfa, , cols) :
   undefined columns selected:   4,   5,   6,   7,   8,   9,  10,  ...

(I think it's not worth the trouble to give the count of missing 
columns, but you might choose to do that.)

2.  Your patch file should be produced by "svn diff", which would have 
included the revision number.

I think your code is an improvement over the current code; let's hope it 
gets accepted.

Duncan Murdoch


On 25/09/2022 2:09 p.m., GILLIBERT, Andre wrote:
> 
> Duncan Murdoch <murdoch.duncan using gmail.com> wrote:
>  > On 24/09/2022 9:56 a.m., GILLIBERT, Andre wrote:
>  > > Dear R developers,
>  > >
>  > >
>  > > One of the error messages that make me loose the most time is the 
> "undefined columns selected" of `[.data.frame`.
>  > >
>  > > It ought to specify the list of bad column names, but currently 
> does not.
>  > >
>  > > Fortunately, this can easily be fixed by a small patch I can write.
>  > >
>  > >
>  > > Are you interested in that patch?
>  >
>  > I doubt if you'll get an answer to this without showing it to us.
>  >
>  > > Is there a standard way to transfer patches for feature requests?
>  >
>  > The standard instructions are here: 
> https://www.r-project.org/bugs.html <https://www.r-project.org/bugs.html> .
>  > You're offering an enhancement rather than a bug fix, but the process is
>  > the same.  But since your change is probably quite small, you might not
>  > need to go through the whole process:  proposing it on this mailing list
>  > might be sufficient.
>  >
>  > Duncan Murdoch
> 
> Please, find the patch attached, based on the latest R SVN trunk code.
> In order to avoid any accidental modification of the behavior of 
> `[.data.frame`, I only inserted code at positions where the "undefined 
> columns selected" error was generated.
> In order to test all code paths, I wrote a small test script (code 
> attached), but did not insert clean & standardized tests to the R build.
> 
> I added a new base::.undefined_columns_msg closure. I am not sure that 
> adding such new internal symbols to the base package is smart because it 
> pollutes the base environment, but looking at the base-internal 
> documentation page, it seems to be standard practice.
> 
> For message translations, I tried to use tools::xgettext() to 
> re-generate the R-base.pot file, but this file seems to be out of sync 
> for a few messages. So, I manually added the new messages to R-base.pot.
> As I am a native french speaker, I added appropriate translations to 
> R-fr.po.
> The make command did not regenerate .mo files, and I did not find any 
> appropriate make target to rebuild them, so, for testing purposes, I 
> manually generated them.  The R installation & administration manual did 
> not help 
> (https://cran.r-project.org/doc/manuals/r-release/R-admin.html#Localization-of-messages 
> <https://cran.r-project.org/doc/manuals/r-release/R-admin.html#Localization-of-messages>).
> 
> When compiling the source code on a x86_64 Ubuntu 16.04 system, I got a 
> problem with position independent code in the internal R LAPACK module. 
> The -fPIC option was not properly passed to the gfortran command line. I 
> fixed that by replacing $(ALL_FCFLAGS) by $(ALL_FFLAGS) on line 16 of 
> src/modules/lapack/Makefile.in
> That looks like a bug to me (misspelling of FFLAGS), but I did not fix 
> that in my patch, since it is a very different issue.
> 
> -- 
> Sincerely
> André GILLIBERT