My Posting on OTN about OWB9.2 new features brought an excellent response from Jean-Pierre Dijcks giving some background on the new release. There's been a couple of follow-ups focussing on the data quality features OWB9.2, together with the first of the new set of Oracle white papers on this area.
A question was raised about whether the new data quality features helped with data quality assessment, such as frequency distributions, looking for child rows with no parents when no FK was ever created, validating numeric data stored as text and so on. From working with other ETL tools such as Ascencial Software Datastage, i've found this to be a useful feature for interactively checking out the source data before building your mapping process.
Nikolai Rochnik (one of the OWB product team) responded to this query which he termed "Data Profiling" and confirmed that, whilst this feature wasn't part of the new data quality features of 9.2, it was slated for the next release of the product. Should be interesting.
Nikolai is also co-author of the first of the OWB9.2 white papers to be made available on OTN, entitled "Oracle Warehouse Builder and Integrated Data Quality". According to the blurb, "This new document describes how Oracle9i Warehouse Builder integrates parsing, correction, data match/merge, and other data cleansing functions with ETL functionality". According to Nikolai, "Name and Address and Match-Merge features in OWB correspond to the Name and Address and Match-Merge operators. However, their capabilities break down further and that's what "other data cleansing functions" refers to. Name and Address = parsing, standardization, correction, augmentation. Match-Merge = de-duplication, householding, record linking."
9:40:59 AM 
|
|
