In an enterprize, merging master data, like customer data, from multiple sources is a common problem. Typically, you do not have a single, i.e. the same key identifying a customer in different sources. You have to match data based on similarity of strings, like names and addresses. In this session, we are going to check how different algorithms for comparing strings included in SQL Server 2008 R2 and SQL Server 2012 (Denali) work. We are going to use Soundex Transact-SQL function, four different algorithms that come in R2 with Master Data Services (Levenshtein, Jaccard, Jaro-Winkler and Ratcliff-Obershelp), and Fuzzy Lookup transformation from Integration Services. Finally, we are going to introduce how SQL Server 2012 Data Quality Services (DQS) help us here. We are also going to tackle the performance problems with string matching merging.

(no tags)