Skip to main content

*** Archive *** Getting best results from Database Locator

Article # 3024354 - Page views: 216


Question / Problem: 

The use of a Associate Search Database (ASD), consisting of a single column, may exhibit unexpected results in regards to the matches and confidence levels.

The use of an ASD in the Transformation Designer may need to be reconsidered.

How to get the best results from the use of a Database Locator?

Answer / Solution: 

The recommended approach is to use either a Local Fuzzy Search database with an increased minimum confidence, or to use a Format Locator with a dictionary and exact match to achieve the required results.

When using a fuzzy database:
It is in the nature of things that in case of faulty input, e.g. by OCR errors, by missing single characters, by swapping characters, entries will still be found with high confidence (of course this makes all the more sense with multiple columns).

A fuzzy database that is used to find exact entries contradicts itself slightly.
Having a database with only one single column is not very common - a Format Locator with a dictionary and an exact search could also be used.

But in the end you can achieve an exact search here too with some configuration:
Use a Fuzzy Database as database.
Include in the .csv file a few numbers that appear exactly the same on the documents and set the minimum confidence of the locator to e.g. 99.5%.
Then only the exact matches will be left.

The option to set the Format Locator for exact matching can be seen in the following image:


The above information is our recommended approach for getting the best results.

Applies to:  

Product Version
KTA 7.7