[rkward-devel] object recovery

[Thomas Friedrichsmeier]

Hi,

On Thursday 22 March 2012, meik michalke wrote:
> if you remove columns from a data.frame (e.g., by overwriting it with an
> empty data.frame) which is opened in RKWard's data editor at the same
> time, you are asked if you would like to recover that data. try this:

[...]

> i don't know how RKWard handles this, but i think it would be best if you
> were not asked one variable at a time, but once for the whole object,
> presenting some kind of list of variables which were removed and could be
> checked to have them recovered. but at least a "don't bother me gain"
> checkbox or a "say 'yes' to all changes to this object" button would be
> nice ;-)

true, I've run into this, before, myself, but shrugged it off, at the time. 
Thinking about it, however, I don't think I ever actually had a reason to 
recover an object. Also, in fact, in the case of rewriting a whole data.frame 
from scratch, the recover functionality will sometimes fail, rather miserably. 
Try this:

> a <- data.frame (a=1:4, b=1:4)
> a
  a b
1 1 1
2 2 2
3 3 3
4 4 4
> rk.edit (a)
NULL
> a <- data.frame (a=1:3, c=1:3)
> # Select "Yes" when asked to restore a$b
> a
  a c
1 1 1
2 2 2
3 3 3
>

See the command log for why this fails.

Perhaps both the easiest and best solution would be to offer recovery for top-
level objects, only, i.e. offer to recover "a", if "a" is found to be removed 
or is no longer a data.frame, but don't offer to recover a[["b"]] as above. 
Probably accidental removal / overwriting of "a" is much more likely than 
accidental removal of a$b, in the first place.

What do you think?

Regards
Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20120325/6b20981f/attachment.sig>