Posted by Nel on 10/10/05 19:58

I am trying to understand the best way to extract a list of users from a
table based upon their location.

To do this I have a table containing all the UK postcodes with a grid
reference x & y.
postcode x y
AB10 392900 804900
AB11 394500 805300
AB12 393300 801100
AB13 385600 801900
AB14 383600 801100
AB15 390000 805300
AB16 390600 807800
AB21 387900 813200
AB22 392800 810700
AB23 394700 813500
AB25 393200 806900
AB30 370900 772900

In a seperate table I have a list of users including a postcode.
id firstname postcode
1 Heather BH4
2 Vicky OL1
3 Paola CF8
4 Joanna W14
5 Steve BD13
6 Sally NN1

In short I would like your opinoin on how best (most efficiently) to
calculate the nearest 20 users. The method I am using to calculate the
distance between the users and any chosen postcode is good old Pythagoras
dist = sq root ( (userX * userX) + (userY * userY) - (locationX * locationX)
+ (locationY * locationY) )

It seems that looping through the users to first calculate the distance is
essential. After that point I am not sure about the besr way to proceed.
Do I write the results to a temp table (with distances) then call them in
order of distance or is there a better way?

Thanks Nel.

[Back to original message]