First Dynamic Data Vault Model Adjustment Achieved!

just what is this?  you heard right…  i’ve been able to absorb changes made to source system data models, and propogate them through to the right place in the data vault – automatically.  now that’s a break-through.  i’ve been working on this for quite a while, and finally achieved it tonight.  it’s not complete, nor is it ready yet for prime-time, but it’s working – and from here it will only improve.

the problem?

you’ve defined a target data vault, and someone gives you an updated source data model with new attributes/elements…  not just one, but hundreds of changes.  perhaps they added comments, perhaps they changed the datatypes, lengths or precisions of some columns, perhaps they added or deleted columns…  now it’s up to you (usually by hand) to modify the target data vault model… 

that could take you a couple days, or a couple weeks to roll out.

the solution…

i’ve got it working now, to automatically push the changes through to the right place for you – takes only a few seconds to achieve….  ok, there’s a few things behind it that make it work (that’s the magic i can’t share)….  but it works.

i believe this is a major breakthrough on the way to “dynamic adjustments to the target data model” be it data vault target, or star-schema target, or even ods (oltp/3nf) target. 

dynamic data model change absorbtion is here…  now if i can just come up with a catchy name, anyone out there got a few ideas? 

by the way, if you’re interested in this bit of software, drop me an email – i’ll be happy to talk to you about my plans for it, and see what you have in mind.

dan linstedt

Tags: , ,

4 Responses to “First Dynamic Data Vault Model Adjustment Achieved!”

  1. Roelant Vos 2010/05/19 at 1:04 am #

    Hi Dan,

    Sounds very good! Well, I’m interested in the software on a personal level. Unfortunately I’m not working on a DV project right now. What I’d like to know is what your default choices are on automatically forwarding changes in the model. There are so many different options when there is a change in a source. In the relatively easy example where, say a new attribute is added or changed. Options include:

    – (always automatically) create a new satellite for the new attribute
    – add the attribute to an existing satellite and recalculate history
    – add the attribute to an existing satellite and start building history from there essentially copying the contents back in history

    I guess the last one would be my choice. Without violating your IP, can you shed some light on these type of decisions? I can see how changes in the model itself (cardinality) could ripple through to a Data Vault automatically and hope to see a demo sometime soon 🙂

    Roelant Vos

  2. dlinstedt 2010/05/19 at 4:03 am #

    Hi Roelant,

    The software works with more than just DV… It works with Star Schemas, and ODS models too (3NF basis).

    The nature of controlling model changes is just that – model changes. It is then up to you (the developer) to decide how to deal with the history (the data in the structure). There are configuration options that let you guide the changes either to a specific model component, or to a new model component. In other words, only the model is changed. Adding a column to a table Satellite doesn’t cause the need to re-calc history, it merely populates it from that point forward. Adding a column to a Hub or link – well, that doesn’t happen – what happens is: new hubs and new links MUST be created (they are at a different grain of data).

    Sorry to be vague, but I’m still in the working phases.

    Hope this helps,
    Dan L

  3. Robert Scott 2019/10/24 at 3:22 am #

    Hi Dan —

    Robert Scott here. I was wondering if you ever made any progress on this that you could share with me. I know of a few universities that are asking on how this can be accomplished and would love to get your thoughts.


  1. Tweets that mention First Dynamic Data Vault Model Adjustment Achieved! | Data Vault Modeling & Methodology -- - 2010/05/14

    […] This post was mentioned on Twitter by erikfransen and erikfransen, Daniel Linstedt. Daniel Linstedt said: First Dynamic Data Vault Model Adjustment Achieved, Source chgs propogated to target model automatically! […]

Leave a Reply