Ask the Experts Question and Answer
Meet the Experts
Ask a Question
(Names of individuals and companies will not be used.)
Ask the Experts Home
(Posted August 22, 2003)
If a piece of information traverses several applications before placement in the DW is the data quality of that information more in question than data that arrives directly from the operational source? If so, is there research that supports this?
Adrienne Tannenbaum?s Answer: Yes, it depends on how much tracking and documentation is available showing what happens to the data on the way ? data flows, meta data flows. Common sense ... if no one knows how something got there or where it came from, I would question data quality.
Chuck Kelleys Answer: I dont think so. Each DW field/element/column has a system of record. That system of record is the place that is discovered to hold the most correct data. So if you have several operational sources that have customer, only one will be chosen as the system of record. That doesnt mean that you wont take data from the others, but that the system of record is the driver. I think if you look at what Bill Inmon has written in Building the Data Warehouse, you will find that he discusses the system of record sufficiently to be that authoritative source. Of course, you could probably use anyone from the "Ask the Experts" panel.
Clay Rehms Answer: This is a great question! From a data quality standpoint, every piece of data that comes into the DW must be questioned no matter how many applications it has traversed. So if that piece of data has had more transformations or more "transports," then you are correct in having more suspicion of that data.
ARCHIVE OF QUESTIONS & ANSWERS FOR DATA QUALITY
BACK TO THE LIST OF CATEGORIES