OWB10g Data Quality Features

Mark Van De Wiel and Ali El Kortobi put together a presentation for Oracleworld on the data quality features in Oracle Warehouse Builder 10g. It covers the name and address cleansing and match-merge features, the new debugging facility and the module snapshot feature.

What's particularly interesting about the new features is the match-merge facility, which allows you to set up 'fuzzy logic' rules to determine whether one or more records are actually logically the same, and merge them in to one record for storing in the warehouse. Unlike the name and address cleansing feature, it doesn't require external data files from companies such as Trillium, and from looking at the paper, seems fairly intuitive to set up.

From what I actually understand, some of these features were actually introduced with OWB9.2 which is already available for download from OTN, so if you're interested in trying out the new features, take a look at this version for a preview of what's coming with 10g.