What kinds of advantages we may have for the integrated use of Stata with DBMS (such as SQL Server)?

[Tatsuru Kikuchi]

What kinds of advantages we may have for the integrated use of Stata with DBMS?

Any advantages for the integration of Stata with DBMS?

I have been much familiar with Database Management System (DBMS) such as SQL Server although I know the Stata plays both a role as a database management system and a role of statistical analysis tool. I would venture to ask a question. What kinds of advantages we may have for the integrated use of Stata with DBMS?

• How about the improvement in speed of analysis in Stata if we use DBMS together?
• How about the improvement in visibility of a whole structure of all data tables?

If any advantages exist, I would like to try to use both Stata and a DBMS together.

Please kindly give me any answers or comments to me!

[Joseph Coveney]

I suspect that you won't be overwhelmed by an improvement in speed of analysis with reasonably small sets of data, although most data management activities execute more quickly in a logically well-designed and maintained relational database than by Stata and this becomes more noticeable as the size of the data grows larger. Visibility of whole structure might be enhanced somewhat, especially with graphical user interface applications that some commercial relational database management systems have, but its practical benefit might not be large for the end user who's just retrieving subsets of the data into Stata for statistical summarization and analysis.

The benefits of working from a formal database are more likely those that are customarily mentioned, such as security (compliance with governmental or organizational policies for restriction of user access, control of changes to the data, data preservation), data-integrity rules (entity integrity, referential integrity, domain integrity, so-called business rules) that you can enforce with commercial relational database management systems, and so on.

To me, the downsides are that creating a database requires planning (there are a lot of design considerations), and that you have to babysit it. You can easily underestimate the amount of effort that these require, and the nature of some projects might not warrant that kind of investment.