My example of merging data was a scenario where the old system was being discontinued and the target system had to do the job of both. That was why it...

• This reply was modified 5 years, 4 months ago by  nova.

Suppose I have to merge data from two systems which happen to have conflicting surrogate primary key values even though they represent...

Roger, I agree with nearly all you have said but when I said before that natural keys could be unique but mutable you said you "vehemently" disagreed with me! Perhaps...

• This reply was modified 5 years, 4 months ago by  nova.
• This reply was modified 5 years, 4 months ago by  nova.

You already said, and I agree, that login names can change. Similarly customer account numbers, so those are two relatively common cases of keys that may need to be updated....

• This reply was modified 5 years, 4 months ago by  nova.

As I said previously however, I am talking about keys in general, not just "primary" keys. A table can have several keys enforced as irreducibly unique and non-nullable but as...

As a developer I vehemently disagree. You need a primary key to be mechanically useful. That means it must be immutable and have other desired characteristics, like uniqueness and...

My point was about keys in general, not "primary"...

Agreed that login name isn't immutable...

I have a question for all you folks claiming that every table must have a natural key.  What would you use for a natural key for either an Employee...

Certainly there should be and it's explicit and implicit throughout...

Chris Harshman wrote:

https://www.kimballgroup.com/2012/02/design-tip-142-building-bridges/

I'm not impressed with the suggestion in that article. Rolling data up into XML and then unwrapping it just to generate a new surrogate key! It's a good thing...

• This reply was modified 5 years, 4 months ago by  nova.

Such a table certainly could contain measures. The concept of a "bridge table" (AKA join table or association table among other names) is actually pretty subjective and difficult to pin...

Have you looked at Dynamic Data Masking in SQL Server?

https://docs.microsoft.com/en-us/sql/relational-databases/security/dynamic-data-masking?view=sql-server-2016

Not sure what you mean by "custom built scripts". You don't need to write any extra scripts to use...

Vinay,

Have you compared the execution plans in SQL Server 2008 and 2016? Is the plan the same?

There are some potential advantages of Python:

• Pandas and NumPy are designed for complex numerical calculations, vectors, matrix manipulations, etc. If you use those features you could very likely find...