Thanks for the advice GSquared, I really appreciate your time answering my question.

This is very useful. Thank you.

Thinking about this more, I wonder if the following will work.

Street names in the UK tend to end with words like Street, Avenue, Crescent etc, if I replace these with...

GSquared, everything you say is correct.

I'm in the UK and yes we do use software to clean addresses for some customer data but others don't actually want their addresses rebuilding...

I've seperated the street number but if anyone can help me remove the redundant information from the address1 I'd be very grateful.

update a

set a.streetno = F.StreetNo

from dbo.test as a

cross...

The more I think about this the more I come to realise that the address fields must be standardised before any merge work is done. I've spent hours tweaking my...

Hi Chris,

I've been trying to work out a stragey for my merging but I'm struggling a little. :blush:

The merge stored procedure will accept two parameters, the source table and the...

Northwest UK yes 🙂 I will PM you.. seriously you deserve drinks for all this help!!!!!

I've made the change and the function is working better now. 🙂

Going to play...

I see.

Another question if you don't mind (you're going to hate me once this little project of mine is done! but if you're anywhere near the northwest then I'd be...

Thanks Chris.

Just to confirm. Is the FuzzyMatch_iTVF2k5 is a full implementation of http://research.microsoft.com/pubs/75996/bm_sigmod03.pdf ?

Hi Chris, I have some questions below. Any advice would be much appreciated. In your original post you said:

Try it out. There's always an element of tuning with this type...

Thanks Chris. You're my new best friend 🙂

I'm going to spend the rest of today on this and will report back in the afternoon.

Once again, I really appreciate your efforts!

Thanks for this Chris but when I tried to create the function I get the following error message:

Msg 156, Level 15, State 1, Procedure FuzzyMatch_iTVF, Line 22

Incorrect syntax near the...

Thanks for this Sean. You are right about the badly written query but for now I'm trying to work out if the approach adopted by the previous developer is the...

Data sets will vary but on average we're talking 2-4 million records.