Re: [O] [BUG] org-table: time difference delivers wrong results

[Nick Dokos]

Martyn Jago <address@hidden> wrote:

> Martyn Jago <address@hidden> writes:
> 
> Apologies I had a typo - please ignore the previous mail. So this looks more 
> sensible:
> 
> > Daniel Bausch <address@hidden> writes:
> >
> >> Anyone?  It might be a duplicate of what Gustav Wiktrom reported on 1st of 
> >> September, but his problem was not resolved, either.
> >>
> >> Daniel
> >>
> >> Am Dienstag 15. November 2011, 22:12:45 schrieb Daniel Bausch:
> >>> Hello list,
> >>> 
> >>> consider the following example:
> >>> |        A |         B |         C |
> >>> |
> >>> |----------+-----------+-----------|
> >>> |
> >>> | 07:11:00 | -08:00:00 | -01:11:00 |
> >>> 
> >>> #+TBLFM: $3=$1+$2;T
> >
> > [...]
> 
> 
> 
> 
> |   Data 1 |   Data 2 |     Total |
> |----------+----------+-----------|
> | 00:00:00 | 00:00:00 |  00:00:00 |
> | 00:00:00 | 00:00:01 | -01:59:59 |
> | 00:00:00 | 00:00:02 | -01:59:58 |
> | 00:00:00 | 00:00:03 | -01:59:57 |
> | 00:00:00 | 00:00:04 | -01:59:56 |
> | 00:00:00 | 00:00:05 | -01:59:55 |
> | 00:00:00 | 00:00:06 | -01:59:54 |
> | 00:00:00 | 00:00:07 | -01:59:53 |
> | 00:00:00 | 00:00:08 | -01:59:52 |
> | 00:00:00 | 00:00:09 | -01:59:51 |
> | 00:00:00 | 00:00:10 | -01:59:50 |
> | 00:00:00 | 00:00:11 | -01:59:49 |
> | 00:00:00 | 00:00:12 | -01:59:48 |
> | 00:00:00 | 00:00:13 | -01:59:47 |
> | 00:00:00 | 00:00:14 | -01:59:46 |
> | 00:00:00 | 00:00:15 | -01:59:45 |
> #+TBLFM: 
> @2$3=$1-$2;T::@3$3=$1-$2;T::@4$3=$1-$2;T::@5$3=$1-$2;T::@6$3=$1-$2;T::@7$3=$1-$2;T::@8$3=$1-$2;T::@9$3=$1-$2;T::@10$3=$1-$2;T::@11$3=$1-$2;T::@12$3=$1-$2;T::@13$3=$1-$2;T::@14$3=$1-$2;T::@15$3=$1-$2;T::@16$3=$1-$2;T::@17$3=$1-$2;T
> 
> |   Data 1 |   Data 2 |     Total |
> |----------+----------+-----------|
> | 00:00:00 | 00:00:16 | -01:59:44 |
> | 00:00:00 | 00:00:17 | -01:59:43 |
> | 00:00:00 | 00:00:18 | -01:59:42 |
> | 00:00:00 | 00:00:19 | -01:59:41 |
> | 00:00:00 | 00:00:20 | -01:59:40 |
> | 00:00:00 | 00:00:21 | -01:59:39 |
> | 00:00:00 | 00:00:22 | -01:59:38 |
> | 00:00:00 | 00:00:23 | -01:59:37 |
> | 00:00:00 | 00:00:24 | -01:59:36 |
> | 00:00:00 | 00:00:25 | -01:59:35 |
> | 00:00:00 | 00:00:26 | -01:59:34 |
> | 00:00:00 | 00:00:27 | -01:59:33 |
> | 00:00:00 | 00:00:28 | -01:59:32 |
> | 00:00:00 | 00:00:29 | -01:59:31 |
> | 00:00:00 | 00:00:30 | -01:59:30 |
> | 00:00:00 | 00:00:31 | -01:59:29 |
> #+TBLFM: 
> @2$3=$1-$2;T::@3$3=$1-$2;T::@4$3=$1-$2;T::@5$3=$1-$2;T::@6$3=$1-$2;T::@7$3=$1-$2;T::@8$3=$1-$2;T::@9$3=$1-$2;T::@10$3=$1-$2;T::@11$3=$1-$2;T::@12$3=$1-$2;T::@13$3=$1-$2;T::@14$3=$1-$2;T::@15$3=$1-$2;T::@16$3=$1-$2;T::@17$3=$1-$2;T
> 
> |   Data 1 |   Data 2 |     Total |
> |----------+----------+-----------|
> | 00:00:00 | 00:00:59 | -01:59:01 |
> | 00:00:00 | 00:01:00 | -01:59:00 |
> | 00:00:00 | 00:01:01 | -01:58:59 |
> | 00:00:00 | 00:01:02 | -01:58:58 |
> | 00:00:00 | 00:01:03 | -01:58:57 |
> | 00:00:00 | 00:01:04 | -01:58:56 |
> | 00:00:00 | 00:01:05 | -01:58:55 |
> | 00:00:00 | 00:01:06 | -01:58:54 |
> | 00:00:00 | 00:01:07 | -01:58:53 |
> | 00:00:00 | 00:01:08 | -01:58:52 |
> | 00:00:00 | 00:01:09 | -01:58:51 |
> | 00:00:00 | 00:01:10 | -01:58:50 |
> | 00:00:00 | 00:01:11 | -01:58:49 |
> | 00:00:00 | 00:01:12 | -01:58:48 |
> | 00:00:00 | 00:01:13 | -01:58:47 |
> | 00:00:00 | 00:01:14 | -01:58:46 |
> #+TBLFM: 
> @2$3=$1-$2;T::@3$3=$1-$2;T::@4$3=$1-$2;T::@5$3=$1-$2;T::@6$3=$1-$2;T::@7$3=$1-$2;T::@8$3=$1-$2;T::@9$3=$1-$2;T::@10$3=$1-$2;T::@11$3=$1-$2;T::@12$3=$1-$2;T::@13$3=$1-$2;T::@14$3=$1-$2;T::@15$3=$1-$2;T::@16$3=$1-$2;T::@17$3=$1-$2;T
> 
> Best, Martyn
> 
> --
> 
> I _will_ engage brain before posting to the ML :(
> 
> 

I can't see straight at this time of the night, but aren't all these formulas 
equivalent to
a column formula?

#+TBLFM: $3 = $1 - $2;T

Assuming that's the case, the problem is that org-table-time-to-string assumes 
that it is given
a positive number and mishandles negative numbers, eg. the first row in the 
last table invokes:

,----
| (org-table-time-seconds-to-string -59 nil)
| "-01:59:01"
`----


I didn't check carefully, but it looks as if the other-way function,
org-table-time-string-to-seconds, deals correctly with "negative" time
strings.

Nick