Arthur Fuller
artful at rogers.com
Wed Oct 22 20:49:14 CDT 2003
Just out of curiosity, has anyone followed this up with actual benchmarks? (I ask this because a man with whom I communicate frequently, Joe Celko, a well-known SQL expert, has said that the lookup speed difference between a 10-char value and a long int on modern systems is the least significant of our problems.) Assuming you code 5+4 zips as 9-digit zips (eliminating the "-"), does anyone on the list have a sufficiently large sample to actually benchmark this meaningfully? By that I mean that about 1M rows, two columns, one of each data type, indexes to suit, and a timer to verify the results? And while I'm on the subject, does anyone have any ideas on how to negate the cache in repeated searches? I'm thinking that the way is to search for non-existent values, but that's just an armchair guess unsupported by facts or knowledge about cache-algorithms. Anybody got such a db that you can send me without violating copyrights etc.? Assuming factorial(9) is the maximum number of 5+4 zip codes, what is the actual number currently? Is there any zip code that begins with "0"? And if so, what about "00"? Is there any logic to the expansion from 5 to 5+4? I.e., they put in a housing development in zip 97600, each of ten buildings sufficiently large to warrant its own zip code, how are they assigned? Incrementally? (Not necessarily step 1). One final question: is there any extant zip code that spans more than one city or town or other regional designation? Curiosity killed the programmer. TIA, Arthur --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.525 / Virus Database: 322 - Release Date: 10/9/2003