[OSM-dev] Overpass API questions

[Alan Mintz]

At 2012-04-27 09:37, Roland Olbricht wrote:
>Alan Mintz wrote:
>
> > Another problem appears to be that [key!=value] filters out any way that
> > does not have a key tag at all, in addition to those that have key=value,
> > which is not what I need.
>
>Yes, this has been the behaviour expected by design. The rationale behind 
>this
>is to filter for "oneway" but not "oneway=no".
>
> > *The point of this is to be able to eliminate ways and their nodes that
> > account for up to 90% of the data being downloaded (i.e. inflating the data
> > by 1000%) when they are unnecessary for a particular editing session.
>
>However, this is a convincing argument. For this reason, and because the
>negation has not found widespread usage so far, I will change the 
>semantics of
>"!=" in the very next version 0.6.98. Version 0.6.98 is scheduled to be
>released on Monday, 30 April.
>
>This means, from 30 April on, you can get with the above query what you
>intended to get.

Wow - that's great. I was going to suggest maybe an alternate syntax, so 
either could be done. Like maybe:
[not(key=value)] would include only ways without key=value and
[key!=value] would continue to include only ways with "key" tags, and their 
values not equal to "value"


--
Alan Mintz <Alan_Mintz+OSM at Earthlink.net>