Matching Mispellings

June 30, 200322 yr

(updated with new attachment 2 posts down)

Hi,

Control entry or duplicates on Names. This method allows to match errors when processing some Name entries. Associated with a Clairvoyance tool, there's little errors that wouldn't be catched IMO :

Match :

- characters inversions

- characters substitution(s)

- character addition(s)

- character ommission(s)

Thanks to Ernst for his help on this one. Originated from this post.

Spell Checker, does this exist ?

Anyone willing to look at the calcs and simplify them to be even more dynamic is really welcome.

Thanks.

SpellCheckerv3.fp5.zip

June 30, 200322 yr

Hey Ugo,

Just had a look at your spellchecker, and I think I'd rather call it 'Fuzzy Clearvoyance tool'! Very nice indeed.

I had a look on the calculations and they looked like rocket science to me, specially the 2,5,11, 23, 47.... sequences...

May I give two suggestions though?...

1. It would be nice if a completely correct entry would become bold or highlighted. I did this in my copy with one extra calc field that compares 'g_name' and 'CustomerName'.

2. It would be good to see clearvoyance results besides the fuzzy 'almost correct' results.

And a little nerdy comment (kick me if I'm wrong!)

All the

Middle("ABCDEFGHIJKLMNOPQRSTUVWXYZ", 1, 1)

Middle("ABCDEFGHIJKLMNOPQRSTUVWXYZ", 2, 1) etc. statements

could really be replaced by

A

B

etc.

Regards,

Ernst

June 30, 200322 yr

Author

Ernst,

Your wish are orders

Here's version 4 with :

- Clairvoyance added at right side

- Clairvoyance added at left side with calculation :

Left(Trim_g_Name;Int(Round(

Lenght(Trim_g_Name)*.65,1))) ---> 65% (one of these intuitive calc I won't explain)...

- Highlight Total Match + Message covering the Portal.

About your comments :

1. Values 2,5,11, 23, 47... for A,B,C,D,E,...

My Math lessons are now far away. What could guarantee uniqueness over a text converted in number ?

The logic I'm using is

A = 2

B = (A*2)+1

C = (B*2)+1

2. Middle("ABCDEFGHIJKLMNOPQRSTUVWXYZ", 1, 1)

Middle("ABCDEFGHIJKLMNOPQRSTUVWXYZ", 2, 1)

For sure, this is a residual from the first test, where each value was calculated immediately as stated above. I'm not sure yet there wouldn't be a easier way, using these Middle("...")*Number for text extracted.

But it's true the result is the same. May be the option you propose would quicken the calc...

For v5