[OSM-dev] extracting house-number from string

[Karl Newman]

On Sat, Mar 14, 2009 at 7:47 AM, Frederik Ramm <frederik at remote.org> wrote:

> Hi,
>
> Stefan Breunig wrote:
> > True about \d. But this regex doesn't match any housenumber in "3rd
> > street". It matches
> > [number][maybe a single letter][word boundary].
>
> Could someone from the US shed some light on this type of address:
>
> 5325 E. Pacific Coast Hwy
> Long Beach, CA 90804
>
> The "E." is, to the best of my knowledge, just a designation of which
> side of the road the building is on; it is neither part of the street
> name (the street is "Pacific Coast Highway", not "East Pacific Coast
> Highway"), nor is it a part of the house number (there is no
> corresponding 5325 A, B, C, D, nor a 5325 W for that matter).
>
> How do you normally treat the "E." when recording addresses in databases
> and so on? (Please don't say "we have a middle initial for addresses" ;-)
>
> Bye
> Frederik
>

Well, it's a bit fuzzy to me, but one thing I'm pretty sure of is that the E
does not designate which side of the road it's on. Almost universally in the
US, odd numbers are on one side, even numbers on the other. The E stands for
East and can mean a few things, based on context. For an address, it's
generally an indication of what geographic area of the highway is being
referred to. In another context (i.e., driving directions), it might mean
the direction of travel on the highway (i.e., Eastbound), although for the
Pacific Coast Highway, I don't think that wouldn't make sense because it's
predominantly a North/South highway. Back to the address: I haven't seen it
so much with named highways, but with numbered streets in certain cities
(i.e., Portland, Oregon) I've seen numbered streets increasing in number in
both directions from a central point in the city, so there could be a E 23rd
Avenue and a W 23rd Avenue that run parallel to each other on different
sides of the central point, so they could be miles apart and you'd better
make sure you know which one you want. In that case the E and W are part of
the street name. Like I said, though, I don't think there's more than one
Pacific Coast Highway so that meaning probably doesn't apply to that
example. The other, probably more common usage that I've seen is that the
East and West refer to different sections of the same street, but the
address numbers increase in both directions from some origin. However, I
think in that case the direction usually shows up as a suffix (i.e., 5325
Pacific Coast Highway E). It's unfortunate that we lost the "Direction"
field from the original TIGER data (it just got munged into the name field)
because then we could probably distinguish between streets with directional
indicators as part of their name (like the E 23rd and W 23rd I mentioned)
and geographical sections of the same street.

Well, now that I've muddied the waters a lot, I'll sign off...

Karl
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20090314/29c5f3b3/attachment.html>