Search masks can be used if a database record contains field that are not always present on a document. You can add a mask so even though some fields are missing on the document, you can still get confident database matches.

For example, an insurance company uses its customer database for several health insurance forms. None of the forms contain all database fields on a single document, yet a customer could be safely identified by one of the following combinations of database fields:

• First name, last name, insurance number

• First name, last name, date of birth

• First name, last name, street, zip code, city

A search mask is a subset of the available database fields. If a document contains only the fields in one of the search masks, it still provides as a viable alternative with a relatively high confidence. Without a search mask, that database entry then is ignored because the document does not enough fields.

When a search mask is used the database locator performs two steps to locate matches on a document.

1. All words in the specified regions of the page(s) are used to find matches in all selected columns from the database configuration.

2. The confidence of each match is calculated by all search masks based on additional parameters in the locator.

Because of this, if you are using a database for multiple purposes such as a source for a database locator and also for a database lookup on the Validation form, Tungsten Automation recommends that you add separate fuzzy databases that use the same source, but with different search columns. This ensures that the database you are using with a database search mask includes only those columns needed for a particular purpose. This also means that you have several fuzzy databases that use the same source data with different column configurations.

Related topics:

• Database Locator

• Set up extraction