[GRASS5] Re: [GRASSLIST:4610] Re: Use of -i option in m.in.e00

[Michel Wurtz]

Radim Blazek wrote:
> (I switched to devel list.)
> 
> On Tuesday 01 October 2002 08:20 pm, Michel Wurtz wrote:
> 
>>The first implementation of m.in.e00 was using feature-ID, and I agree with
>>you. The problem comes from the info tables, where feature-# is used for
>>linking graphic objects and attributes.  So you can have multiple entries
>>for one feature-ID (where some fields, like perimeter or surface,
>>different). It is then safer to use feature-#,
> 
> 
> Yes.
> 
> 
>>except when the feature-# is
>>just what you want in dig_att (like Z value of contour lines...)
> 
> 
> This i don't understand. I would understand: "except when the feature-ID is 
> ...". 

Ooops, you're right. I thinked feature-ID and wrote feature-# ...must be tired :-(

> How is the feature-ID used in Arc/Info if it is not the link to attributes?
> Are attributes (other then surface etc.) synchronized somehow 
> if two objects have the same feature-ID?

feature-ID is an "external" identifier, which remind after saving and restoring
data, cleaning, etc... So it is used to link an Arc/Info coverage to some
external table (RDBMS, Excel, etc...)
feature-# in an "internal" identifier used to ensure the coherence between the
Arc (graphic) and the Info (attribute tables) parts.

Practicaly, using same feature-ID for different objects is not common.
It can be used for polygons with islands, when you don't want an unique object,
because you want to be able to display the surface of each island, but ensure
that all other attributes are the same.
There is probably other examples...

Michel.