[GRASSLIST:5000] Re: [GRASS5] v.patch in 5.7 problems

>> Maciek Sieczka wrote:
>>
>>> grass57_exp_2004_11_13
>>>
>>> After merging a vector consisted of points with a vector consisted of
>>> lines using v.patch I get a file where some lines and points are
>>> assigned the same "field:" and "category:". It unables me create a
>>> database which would assign the elevation values to both points and
>>> contour lines at one time! And I need it to be done so in order to
>>> interpolate this "merged" elevation information in one turn with
>>> v.surf.rst. Both input files had a nice database.
>>>
>>> BTW - why doesn't the v.patch create a merged database for the output
>>> file while in 5.5 the v.patch copies all the input's cats and labels
>>> to the output?
>>>
>>> Please somebody answer and let me know if it's a bug worthy of a bug
>>> report.
>>
>>
>> It is wish report.
>
>
> Thanks for the reply.
>
> After re-thinking my problem I'd say that it should rather be a regular
> bug - in my opinion in one vector file there may not be two diffrent
> features having the same "field:" and "category:" which happens when
> using v.patch. Am I wrong?

Both can be useful, preserve cats or assign new unique cats.
Ideally v.patch should support both. A Bug is if it does not do what it
should - what is written in man. A wish is if a feature is missing.

According to what you say I still claim that the issue that v.patch doesn't assign unique cats to each single element of a resulting vector file is a bug, since such a feature *is present* in v.patch in Grass 5.0/5.5. Not porting this feature to v.patch in 5.7 is an errorous disabling one of it's features.

Maciek

Maciek Sieczka wrote:

After re-thinking my problem I'd say that it should rather be a regular bug - in my opinion in one vector file there may not be two diffrent features having the same "field:" and "category:" which happens when using v.patch. Am I wrong?

Both can be useful, preserve cats or assign new unique cats.
Ideally v.patch should support both. A Bug is if it does not do what it
should - what is written in man. A wish is if a feature is missing.

According to what you say I still claim that the issue that v.patch doesn't assign unique cats to each single element of a resulting vector file is a bug, since such a feature *is present* in v.patch in Grass 5.0/5.5. Not porting this feature to v.patch in 5.7 is an errorous disabling one of it's features.

Send us a patch.

Radim

PS:
1. don't remember to check if all tables have the same structure
    (column names and types)
2. Changing v.patch from low level (geometry elements) to high level
    (features with attributes) means that you have to overlay areas,
     i.e. to do something similar like v.overlay, otherwise
     it can work only on points and lines - do not forget
     that one geometry can represent more features.