[R] R is not for Relational

Hoon Yoon - IPT Quant hyoon at nyptsrv1.etsd.ml.com
Wed Aug 4 14:52:08 CEST 1999


SQL could do a lot more than just merge/match/subset can ever do in a single 
statement. I think key here is that it can do so much in a single statement like 
using DISTINCT, HAVING, etc...
Given that R does not have to loop is this major improvement? In terns of being 
able to accomplish alot with single statment and retain clearity. I think so. 
Can people live without it? Certainly we can. Unlike C, we do not have to loop 
to get stuff like this done through memory. (Although some might argue that you 
can call SQL from C, I would say you have to have a DB engine and that means 
dumping it to disk, which is a slow operation.) Do I think it's hell of 
improvement? Sure it is. Nice interpreted lang like R is used for this reason. 
Ease and readability.
Is this a easy quick and dirty job. Not likely.
I think the author should look at gadfly, which is Python (another language) in 
memory SQL database. It's open source and Python is easy lang. Although some 
uses this, I have to caution that it's slow for any complicated query. And Aaron 
has his PhD in Relational database.
Still, I think just subset of mult-table join would prove very valuable. Love to 
 see it done.
Just my 2cents.


r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list