[GRASS-user] populating dbtable with value sequence

[Moritz Lennert]

On 18/04/13 18:16, Shaun Langley wrote:
>
> On Apr 18, 2013, at 9:48 AM, Moritz
> Lennert<mlennert at club.worldonline.be>  wrote:
>
>> On 18/04/13 15:52, Shaun Langley wrote:
>>> **This question is also posted on stackexchange.com
>>> <http://stackexchange.com>**
>>>
>>> I have a vector map that includes 100 points. These points were
>>> created using |v.extract| to randomly sample a larger map of 10k
>>> points. As such, the 'cat' values range from 1 to more than 10k.
>>> What I would like to do is to update the attribute table by
>>> rescaling these values to 1 to 100.
>>>
>>> Here's what I've experienced. If I use
>>>
>>> |v.category input=pts100 output=foo option=del v.category
>>> input=foo output=pts100 option=add --o|
>>>
>>> Then when the 'cat' values /appear/ rescaled when I print them
>>> with
>>>
>>> |v.category input=pts100 option=print|
>>>
>>> However, if I look at the dbtable with
>>>
>>> |db.select pts100|
>>>
>>> the 'cat' values have not changed (presumably because it is the
>>> key column).
>>>
>>> So I'm thinking I will need to add a new column using
>>> |v.db.addcol|, but I don't know how to specify the range of value
>>> to populate the new column. From what I can tell, I have to
>>> specify a single value.
>>>
>>> Ideas?
>>
>> Either
>>
>> - v.distance (possibly with dmax=0) to get the cat values from the
>> new points into the old map and then v.reclass to use the new
>> values as cat values.
>>
>
> It isn't clear to me how this can be used to rescale the cat values.

I meant using the new vector map pts100 as 'to_map' and your original 
map as 'from_map' and filling in a new column in the original map with 
the cat values of pts100. dmax=0 should work as the points are identical.


> Points map has cat values that are (for example)
>
> 120 343 444 521 1334
>
> I don't necessarily have to rescale the cat values if I can create a
> new column in which the values are scaled 1-100 (it's not feasible to
> do this by hand given the quantity of values I'd have to change.  I
> need to script it.)

If you use PostgreSQL as a data backend, you can simply create a 
sequence and fill up the table with calls to nextval().

I don't think SQLite has sequences as such (only autoincrement for rowid's).

Otherwise, I agree that you will probably have to script, so maybe the 
v.category + v.distance is actually the easiest.

Or, just thought of another possible solution (untested): use v.category 
to add a new layer with your ordered cat values to your original map, 
v.db.addtable to add a table to that layer, v.db.addcol to add a new 
column to the original attribute table in layer 1 cand v.db.update with 
qlayer and either option=cat (if this works, you actually don't have to 
create an attribute table for layer 2) or option=query and qcolumn.

Moritz