Re: st: Data Management Tool

[David Kantor <kantor.d@att.net>]

Nick Cox had some valuable responses to these questions.

I would add...
Stata is a RAM-only system; it does not pipeline large datasets. At 
least, as far as I know, this is correct.  Thus, you need large 
amounts of memory to work with large datasets.  If your datasets are 
really huge (and if I'm correct that Stata can't pipeline), then this 
may be a limitation.  I believe that Stata's choice to use an 
all-in-memory method is based on the idea that, while datasets can be 
large, memory keeps getting more cheap and plentiful.

I'm not sure about querying datasets in foreign formats (Oracle, 
mysql).  You usually convert them to Stata before using them. Thus, 
there is a static aspect to that operation.  That is, you don't get 
to query a live system -- one that is subject to changes by other 
users while you are accessing it. You get to access a snapshot of a 
system, as of the moment the data were extracted.

But then, there is the odbc set of facilities.  I'm not familiar with 
them, but they do allow access to some foreign-format data.  (Whether 
it allows access to live systems, I don't know.)

Finally, I do know of one large commercial product for data 
processing & reporting that has been built using Stata as its 
core.  This is produced by Prof-soft Health.  You may want to contact 
the person who put this together; that is Ed Bassin, 
edbassin@prosoft-health.com.

I hope this helps; Good Luck.
--David

At 10:42 AM 2/6/2007, Marc wrote:

> Hi All,
>
> I'm looking into Stata MP as a data management system on linux for 3-5 users.
>
> The main tasks of the data management system would be to query 
> tables from Oracle, mysql, create variables, run statistical 
> analysis/modeling and support fairly advanced reporting.
>
> I've worked with SAS and SPSS, however this time around I'd be the 
> one paying for the license, and am being consciencious about 
> shopping for the best solution. The corporate reflex of "just 
> writting a PO" for SAS or SPSS server is not the chosen approach here.
>
> - How does Stata compare as data management tool, specially as it 
> interfaces with databases?
>
> - Is it a RAM-only system, or does it pipeline large datasets ?
>
> - What are its limitations as a data management tool ?
>
> - At what point would you say SAS/SPSS is hands down the best 
> solution (if at all)?
>
> - Any opinion on its syntax/batch capabilities?
>
> I know next to nothing about Stata's macro capabilities, and one 
> thing I like about SPSS is its new Python extensibility. Anything 
> comparable with Stata ?
>
> Many thanks for your time.
>
> Marc.
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/