Fuzzy matching do not work as expected

[Mette_Foss]

Hi,

I am having this problem with the fuzzy matching, where I cannot find any settings that match the three rows below - I have tried multiple settings, and also the address, zip code, company match etc. - see below.

For the tree rows Kreditornumber, Bogføringsdato, and Beløb I am looking for an exact match - it is for the Ekstern Faktura fuzzy match (Ideally with a 80 pct. threshold) the problems appear - the three lines do not come out as matched rows.

Can anyone help me with the settings that match the three rows? I have more than 80K lines in my dataset, so it is important that the rules are not too weak - because then I will receive too many false positives. 

RecordID	Kreditornumber	Bogføringsdato	Ekstern Faktura	Beløb
1	36050494	22-02-23	AX391059	150.620,25
2	36050494	22-02-23	AX391060	150.620,25
3	36050494	22-02-23	AX391061	150.620,25

Thanks in advance,

Br,

Mette

[lynnesonney · Accepted solution]

@Mette_Foss Fuzzy match is really more of an art than a science. For IDs like your "Ekstern Faktura," I recommend changing the match function to Character oriented rather than words. Also - keep in mind that "Generate Keys" is based on the way text SOUNDS, so this isn't very useful for IDs like we have here. I would recommend trying a setup like THIS  (see images)  on your full data set and evaluating if this meets your needs. 

Be sure to accept this as a solution if it works!

[lynnesonney]

For reference, workflow attached.