Mark,

	Gotcha. I'd suggest reviewing the data I posted a link to.

	As an example, '37642' appears to be 'CHURCH HILL, TN' according to http://zip4.usps.com/zip4/zcl_3_results.jsp 
  and my data indicates it as '37642	TN	Tennessee'. I don't see any of  
these 54 conflicts in, admittedly limited, checking within my data.  
You may just have an issue with the original source data. In fact, the  
USPS site won't accept alphabetic characters like those in '063HH CT  
NY' as input.

	I have more extensive data that includes much more data for each zip  
code, including town names. And it seems to correlate with the USPS  
site data, but this subset is much more manageable than it's 36  
million characters.

P


On Mar 21, 2009, at 6:49 PM, Mark J. Reed wrote:

> On Sat, Mar 21, 2009 at 1:27 PM, Paul Skinner <[log in to unmask]>  
> wrote:
>>        Late to the party, but how was the fact that there are 42 zips
>> crossing state lines determined?
>
> From the Census data file posted way back in the second message.
> http://www.census.gov/tiger/tms/gazetteer/zcta5.txt
>
> 42 was actually conservative, since I was only considering the
> all-numeric ZIP codes.  A more general count reveals 54 overlapping
> zips:
>
> 063HH CT NY
> 380HH AR KY
> 380HH AR TN
> 380HH KY TN
> 422XX KY TN
> 582HH MN ND
> 592HH MT ND
> 657HH AR MO
> 976HH CA OR
> 979HH ID OR
> 10004 NJ NY
> 37642 TN VA
> 38041 AR TN
> 38063 AR TN
> 38079 KY TN
> 38852 AL MS
> 42223 KY TN
> 51630 IA MO
> 51640 IA MO
> 52626 IA MO
> 52761 IA IL
> 55954 IA MN
> 56027 IA MN
> 56129 IA MN
> 56219 MN SD
> 57638 ND SD
> 59221 MT ND
> 65733 AR MO
> 65761 AR MO
> 67950 KS OK
> 69337 NE SD
> 71749 AR LA
> 72644 AR MO
> 73949 OK TX
> 79922 NM TX
> 82063 CO WY
> 82082 NE WY
> 83120 ID WY
> 84536 AZ UT
> 85534 AZ NM
> 86044 AZ UT
> 86504 AZ NM
> 86508 AZ NM
> 86514 AZ NM
> 86514 AZ UT
> 86514 NM UT
> 86515 AZ NM
> 87328 AZ NM
> 89439 CA NV
> 97635 CA OR
> 97910 ID OR
> 97913 ID OR
> 99128 ID WA
> 99362 OR WA
>
>
> -- 
> Mark J. Reed <[log in to unmask]>