R help - Apr 2014 - better an error?

I just spent about an hour bug-tracking.  I had expected the following to
throw an error:

  d <- data.frame( x=1:5, y=6:10 )
  valid <- c(TRUE, FALSE)
  d[valid,]

I understand that R recycles "when fit," but I had not expected it to
recycle, then truncate, and not give even a warning.  maybe there is a good
reason for this.

I would love to be able to teach R to my MFE students.  alas, I don't feel
that I can inflict on them the mysterious errors in R.  this ranges from
poor checking of when variables exist to auto-recycling (without an ability
to turn this off even with an option) to the non-printing of the last
numbered R source code statement upon an error (that I can see in the
traceback()) to non-expected behavior (e.g., subset(d,x,select=-c("a",
"b"))) to .  I know many of these issues can be fixed and/or do not
bother
the experts, and I am personally happy to live with R for its power despite
its drawbacks; but IMHO it is just too much to ask from a set of bewildered
novice master students.

I hope the R team will at some point in the future pick up on making the
core language less mysterious upon setting an option, at least in "user
space".

/iaw

----
Ivo Welch (ivo.welch@gmail.com)

	[[alternative HTML version deleted]]

The behaviour that you get is exactly the behaviour that I, at least, 
would expect, and it seems to me to be exactly the correct behaviour.
I do not understand what you are complaining about.

cheers,

Rolf Turner

On 11/04/14 06:31, ivo welch wrote:
> I just spent about an hour bug-tracking.  I had expected the following to
> throw an error:
>
>    d <- data.frame( x=1:5, y=6:10 )
>    valid <- c(TRUE, FALSE)
>    d[valid,]
>
> I understand that R recycles "when fit," but I had not expected
it to
> recycle, then truncate, and not give even a warning.  maybe there is a good
> reason for this.
>
> I would love to be able to teach R to my MFE students.  alas, I don't
feel
> that I can inflict on them the mysterious errors in R.  this ranges from
> poor checking of when variables exist to auto-recycling (without an ability
> to turn this off even with an option) to the non-printing of the last
> numbered R source code statement upon an error (that I can see in the
> traceback()) to non-expected behavior (e.g.,
subset(d,x,select=-c("a",
> "b"))) to .  I know many of these issues can be fixed and/or do
not bother
> the experts, and I am personally happy to live with R for its power despite
> its drawbacks; but IMHO it is just too much to ask from a set of bewildered
> novice master students.
>
> I hope the R team will at some point in the future pick up on making the
> core language less mysterious upon setting an option, at least in
"user
> space".