Re: [O] Table cell refs with @0 or $0 are broken

[Christian Moe]

On 4/4/12 7:17 PM, Nick Dokos wrote:
> James Harkins<address@hidden>  wrote:
>
> > Don't know if this is fixed in a later org update, but -- the online
> > org manual says that you can refer to table cells in the current row
> > or column using @0 and $0 respectively, but that's definitely not
> > working on my machine.
(...)
> > ** Broken: "Not in table data field"
> > | 1.0 |
> > | 2.0 |
> > |     |
> > #+TBLFM: @>$1=vsum(@address@hidden)
(...)
> Did this ever work? I've spot-checked back to 6.36c and I cannot find
> a release where it actually worked: assuming I haven't made a mistake,
> it seems to be an implementation oversight, rather than some patch
> specifically breaking the functionality.
>
> Nick

I find that zero references do work, albeit not quite as one might be 
led to believe by the manual.

Agreed, it does not work in James' example. And agreed, the manual 
seems to me to suggest, to the contrary, that address@hidden' should work. 
Though it does also make clear that the `$0' is superfluous and can be 
omitted, because when only the row is given, the current column is 
taken as implied. So we all agree address@hidden' works.

But consider the following example, which also works. Here is some 
output from fitting a linear trend in Org-Babel/R. (I've shaved off 
some decimals to fit this in an email.) Computing the TBLFM will round 
the number in each cell to three decimal places.

#+results:
| Record   |        Slope |    ConfLower |    ConfUpper |
|----------+--------------+--------------+--------------|
| GISTEMP  | 0.0173837600 | 0.0133209130 | 0.0214466060 |
| HadCrut3 | 0.0158602890 | 0.0118664610 | 0.0198541180 |
#+TBLFM: @2$2..@>$>address@hidden;%.3f

Try substituting `$0' for address@hidden', it works the same. @0 designates the 
current row, and the current column is taken as implied. Ditto when $0 
designates the current column. However, address@hidden' will not work.

It would seem that either one of $0 and @0 on its own in practice 
designates "the current cell".

Also, I can't think of a situation where either would be needed to 
designate a whole column or row respectively.


Yours,
Christian