[Rd] Bug in prettyNum

[Martin Maechler]

Thank you,  Marttila  and Ivan,

As the original author of prettyNum() {etc ..},
I will commit such a bug fix to R-devel (and probably port it to
R 4.5.0 patched) quite soon
(but not yet today).

Best regards,

Martin Maechler

>>>>> Ivan Krylov via R-devel 
>>>>>     on Fri, 23 May 2025 17:14:57 +0300 writes:

    > В Fri, 23 May 2025 11:47:33 +0000
    > Marttila Mikko via R-devel <r-devel using r-project.org> пишет:

    >> When called with a numeric vector, the `replace.zero` argument is
    >> disregarded.
    >> 
    >> > prettyNum(0, zero.print = "- ", replace.zero = TRUE)  
    >> [1] "-"
    >> Warning message:
    >> In .format.zeros(x, zero.print, replace = replace.zero) :
    >> 'zero.print' is truncated to fit into formatted zeros; consider
    >> 'replace=TRUE'

    >> Please see below a patch which I believe would fix this.

    > Surprisingly, it's not enough. The 'replace' argument to .format.zeros
    > needs to be "threaded" through both the call to vapply(x, format, ...)
    > and the internal call from format.default(...) to prettyNum(...):

    R> options(warn = 2, error = recover)
    R> prettyNum(0, zero.print = "--", replace.zero = TRUE)
    > Error in .format.zeros(x, zero.print, replace = replace.zero) :
    > (converted from warning) 'zero.print' is truncated to fit into formatted zeros; consider 'replace=TRUE'

    > Enter a frame number, or 0 to exit

    > 1: prettyNum(0, zero.print = "--", replace.zero = TRUE)
    > 2: vapply(x, format, "", big.mark = big.mark, big.interval = big.interval, sma
    > 3: FUN(X[[i]], ...)
    > 4: format.default(X[[i]], ...)
    > 5: prettyNum(.Internal(format(x, trim, digits, nsmall, width, 3, na.encode, sc
    > 6: .format.zeros(x, zero.print, replace = replace.zero)
    > 7: warning("'zero.print' is truncated to fit into formatted zeros; consider 'r
    > <...omitted...>
    > Selection: 6
    > <...>
    > Browse[1]> ls.str()
    > i0 :  logi TRUE
    > ind0 :  int 1
    > nc :  int 1
    > nx :  num 0
    > nz :  int 2
    > replace :  logi FALSE
    > warn.non.fitting :  logi TRUE
    > x :  chr "0"
    > zero.print :  chr "--"

    > Since prettyNum() accepts ... and thus ignores unknown arguments, it
    > seems to be safe to forward the ellipsis from format.default() to
    > prettyNum(). The patch survives LANGUAGE=en TZ=UTC make check-devel.

    > Index: src/library/base/R/format.R
    > ===================================================================
    > --- src/library/base/R/format.R	(revision 88229)
    > +++ src/library/base/R/format.R	(working copy)
    > @@ -73,7 +73,7 @@
    > decimal.mark = decimal.mark, input.d.mark = decimal.mark,
    > zero.print = zero.print, drop0trailing = drop0trailing,
    > is.cmplx = is.complex(x),
    > -			     preserve.width = if (trim) "individual" else "common"),
    > +			     preserve.width = if (trim) "individual" else "common", ...),
    > ## all others (for now):
    > stop(gettextf("Found no format() method for class \"%s\"",
    > class(x)), domain = NA))
    > @@ -338,7 +338,8 @@
    > big.mark=big.mark, big.interval=big.interval,
    > small.mark=small.mark, small.interval=small.interval,
    > decimal.mark=decimal.mark, zero.print=zero.print,
    > -		    drop0trailing=drop0trailing, ...)
    > +		    drop0trailing=drop0trailing, replace.zero=replace.zero,
    > +		    ...)
    > }
    > ## be fast in trivial case, when all options have their default, or "match"
    > nMark <- big.mark == "" && small.mark == "" && (notChar || decimal.mark == input.d.mark)


    > -- 
    > Best regards,
    > Ivan

    > ______________________________________________
    > R-devel using r-project.org mailing list
    > https://stat.ethz.ch/mailman/listinfo/r-devel