[bug] removing last column with a column formula

[bug] removing last column with a column formula

[Nicolas Goaziou]

Hello,

Let's consider the following table:

#+begin_src org
| 1 | 2 | 3 |
| 1 | 2 | 3 |
#+TBLFM: @2$1..@2$3=@1
#+end_src

If I remove the second column (M-S-Left), the formula is correctly
updated. But when I remove the last column, the formula gets partly
deleted and becomes:

                            #+TBLFM: @2$1..

Also, if this times, the first column is removed, that line becomes:

                      #+TBLFM: @2$INVALID..@2$2=@1

This is with latest Org and no user configuration.

Regards,

-- 
Nicolas Goaziou

Re: [bug] removing last column with a column formula

[Michael Brand]

Hi Nicolas

Columns and rows are better referenced with "<" and ">" in many cases
to avoid such oddities:

#+begin_src org
| 1 | 2 | 3 |
| 1 | 2 | 3 |
#+TBLFM: @2$<..@2$>=@1
#+end_src

http://orgmode.org/manual/References.html#References

Michael

On Tue, Aug 9, 2011 at 14:50, Nicolas Goaziou <n.goaziou@gmail.com> wrote:
> Hello,
>
> Let's consider the following table:
>
> #+begin_src org
> | 1 | 2 | 3 |
> | 1 | 2 | 3 |
> #+TBLFM: @2$1..@2$3=@1
> #+end_src
>
> If I remove the second column (M-S-Left), the formula is correctly
> updated. But when I remove the last column, the formula gets partly
> deleted and becomes:
>
>                            #+TBLFM: @2$1..
>
> Also, if this times, the first column is removed, that line becomes:
>
>                      #+TBLFM: @2$INVALID..@2$2=@1
>
> This is with latest Org and no user configuration.

Re: [bug] removing last column with a column formula

[Nicolas Goaziou]

Hello,

Michael Brand <michael.ch.brand@gmail.com> writes:

> Columns and rows are better referenced with "<" and ">" in many cases
> to avoid such oddities:
>
> #+begin_src org
> | 1 | 2 | 3 |
> | 1 | 2 | 3 |
> #+TBLFM: @2$<..@2$>=@1
> #+end_src

Thanks for the pointers.

Though, the object of my message was more to point at a strange
behaviour that, perhaps, should be tackled down, someday.

Regards,

-- 
Nicolas Goaziou

Re: [bug] removing last column with a column formula

[Bastien]

Hi Carsten,

Nicolas Goaziou <n.goaziou@gmail.com> writes:

> Let's consider the following table:
>
> #+begin_src org
> | 1 | 2 | 3 |
> | 1 | 2 | 3 |
> #+TBLFM: @2$1..@2$3=@1
> #+end_src
>
> If I remove the second column (M-S-Left), the formula is correctly
> updated. But when I remove the last column, the formula gets partly
> deleted and becomes:
>
>                             #+TBLFM: @2$1..
>
> Also, if this times, the first column is removed, that line becomes:
>
>                       #+TBLFM: @2$INVALID..@2$2=@1
>
> This is with latest Org and no user configuration.

Is this something that we can (easily) fix?  

Thanks!

-- 
 Bastien

Re: [bug] removing last column with a column formula

[Carsten Dominik]

On 24.8.2011, at 20:29, Bastien wrote:

> Hi Carsten,
> 
> Nicolas Goaziou <n.goaziou@gmail.com> writes:
> 
>> Let's consider the following table:
>> 
>> #+begin_src org
>> | 1 | 2 | 3 |
>> | 1 | 2 | 3 |
>> #+TBLFM: @2$1..@2$3=@1
>> #+end_src
>> 
>> If I remove the second column (M-S-Left), the formula is correctly
>> updated. But when I remove the last column, the formula gets partly
>> deleted and becomes:
>> 
>>                            #+TBLFM: @2$1..
>> 
>> Also, if this times, the first column is removed, that line becomes:
>> 
>>                      #+TBLFM: @2$INVALID..@2$2=@1
>> 
>> This is with latest Org and no user configuration.
> 
> Is this something that we can (easily) fix?  


Not easily.  I have made it throw an error now, but there may be
other situations like this which are not yet captured.

Michael Brand has pointed out correctly that if you want to use
ranges the go to the border of the table, you should use
$< and $>

Cheers

- Carsten

Re: [bug] removing last column with a column formula

[Bastien]

Carsten Dominik <dominik@uva.nl> writes:

>> Is this something that we can (easily) fix?  
>
> Not easily.  I have made it throw an error now, but there may be
> other situations like this which are not yet captured.

Throwing an error in this case is already an improvement -- thanks 
for this!

-- 
 Bastien