[Rd] Surprising behavior when using the reference class with the dollar symbol

[Bill Dunlap]

> > It looks like when calling the dollar symbol using the function format, it
> > treats the input argument as a character literal and does not evaluate it
> > inside the function. I know we have the function `field` to get the slot
> > variable, but I wonder if this is designed on purpose as the above example
> > works for both list and S4 system.

It works almost the same way with lists, the difference being that an
unknown name with a list gives NULL and not an error.

> L <- list(A=101)
> `$`(L, A)
[1] 101
> `$`(L, "A")
[1] 101
> i <- "A"
> `$`(L, i)
NULL

The 2nd argument is treated as a symbol and is never evaluated.  If it
is given as a string literal then it is converted to a symbol - that
is a holdover from the old days before backticks could be used to make
symbol literals for symbols that included odd things like spaces and
dollar signs.

-Bill

On Sat, Mar 27, 2021 at 7:49 AM Jiefei Wang <szwjf08 using gmail.com> wrote:
>
> Thanks, Duncan. Below is the repost of my question in plain text mode.
>
>
> I'm trying to get the field value of a reference object by the field
> name, but the dollar symbol behaves quite unusual. See example below
>
> .foo <- setRefClass(
>     "foo",
>     fields = list(
>         a = "integer"
>     )
> )
> x <- .foo$new(a=1L)
> ## This is OK
> x$a
> ## This is OK
> `$`(x, "a")
> ## But this is not OK
> i <- "a"
> `$`(x, i)
>
> For the last line of code I get this error
>
> Error in envRefInferField(x, what, getClass(class(x)), selfEnv) :
>   ‘i’ is not a valid field or method name for reference class “foo”
>
> It looks like when calling the dollar symbol using the function
> format, it treats the input argument as a character literal and does
> not evaluate it inside the function. I know we have the function
> `field` to get the slot variable, but I wonder if this is designed on
> purpose as the above example works for both list and S4 system.
>
> Best,
> Jiefei
>
>
> On Sat, Mar 27, 2021 at 10:20 PM Duncan Murdoch
> <murdoch.duncan using gmail.com> wrote:
> >
> > On 27/03/2021 10:16 a.m., Jiefei Wang wrote:
> > > Hi all,
> > >
> > > I'm trying to get the field value of a reference object by the field name,
> > > but the dollar symbol behaves quite unusual. See example below
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > *.foo <- setRefClass(    "foo",    fields = list(        a = "integer"
> > > ))x <- .foo$new(a=1L)## This is OKx$a## This is OK`$`(x, "a")## But this is
> > > not OKi <- "a"`$`(x, i)*
> >
> > This is really hard to read.  Please post it again, but don't use HTML.
> >
> > Duncan Murdoch
> >
> > >
> > > For the last line of code I get this error
> > >
> > >
> > > *Error in envRefInferField(x, what, getClass(class(x)), selfEnv) :   ‘i’ is
> > > not a valid field or method name for reference class “foo”*
> > >
> > > It looks like when calling the dollar symbol using the function format, it
> > > treats the input argument as a character literal and does not evaluate it
> > > inside the function. I know we have the function `field` to get the slot
> > > variable, but I wonder if this is designed on purpose as the above example
> > > works for both list and S4 system.
> > >
> > > Best,
> > > Jiefei
> > >
> > >       [[alternative HTML version deleted]]
> > >
> > > ______________________________________________
> > > R-devel using r-project.org mailing list
> > > https://stat.ethz.ch/mailman/listinfo/r-devel
> > >
> >
>
> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel