Frequently Asked Questions
Q: So there is no Data Vault 3.0?
A: No there are no 3.0 and there is no 1.0, the Data Vault users does not need version numbers on a modelling methodology. To put a version number on a modelling methodology is just plain stupid. As you probably know there are no Third normal form (3NF) 2.0 or Star Schema 2.0 either.
Q: Is there a Data Vault 1.0?
A: No there have never been a 1.0. Data Vault is called Data Vault even when it has been improved over the years. See the question above.
Q: Why is Data Vault 2.0 called Dan Lindstedts Data Vault 2.0?
A: Data Vault is an ever-evolving methodology and technique but Dan Lindstedt did freeze one flavor in time and claimed ownership and copyrights etc on DV2.0. As he claims sole ownership of this flavor it is Dan Lindstedts Data Vault 2.0.
Q: We use Hash keys but not Data Vault 2.0. Do we need to
upgrade?
A: Not at all, use of Hashkeys is very much recommended
in Data Vault when the use of it is motivated. When not needed or when it risks
complicating the solution it can be avoided.
Q: Do I need DV 2.0 to use Hash keys?
A: No. Hash keys have and should be used in any
Data Vault when the use of it is motivated.
Q: We are using DV2.0 and Link Satellites. I heard
Link-Satellites are outdated and problematic, what should I do?
A: Link Satellites are related to numerous
problems and therefor discouraged to use in any modern Data Vault
implementation. Just stop using them as most Data Vault modelers have done now.
Dan Lindstedts Data Vault 2.0 does use these constructs.
Q: I’m Certified Data Vault Data Modeler, is my certification
relevant and modern still or do I need to Certify in DV2.0?
A: No need. But if it was a while since you
updated your knowledge on Data Vault, please read up on the latest developments
or even take a refresher course that is up to date now in 2017.
Q: We have followed the recommendation in Data Vault 2.0 and
implemented all our transactions as links with a timehub connected to it. We
now have huge problems with it. Is that not correct?
A: Afraid so. The practice of using links to
model transactions has been proven outright wrong and severely pattern-breaking
by most experts. Regardless of how detailed you slice your time hub you will
always run the risk of losing transactions if they occur in the same time frame
since that row in the link already exists. This pattern exists in Data Vault 2.0
though which serves to further illustrate the problems of freezing a point in
time-version of an evolving methodology.