[R] MySQL + R as a Replacement for SAS Proc SQL + Various Stat Procs

[JWilliamson at lecg.com]

For cost reasons, I'd like to replace SAS on my PC under Win XP Pro.

Nearly all my work involves medium-size datasets (100k-10M) records which I
cleanup, relate, fliter and get into shape for analysis using SAS/SQL
followed by standard statistical procedures, e.g. regression using SAS proc
reg.

It seems to me that this type of analysis could be done in MySQL followed
by R, but I'd like some advice about the best way to pass datasets from
MySQL to R.  I understand there are various connectivity packages
avaialble, RODBC and RMySQL, but I'd appreciate some advice about where to
dig in first.

I'd like to avoid additional syntax in my SQL code -- so it seems better
not to coat SQL queries in R wrappers and pass them to MySQL -- probably
better for my way of working to finish all the SQL work and pass a clean
table ready to analyze to R.

One of the great advantages for me using SAS is that I can beat the data
into shape using proc SQL and then call proc STAT_OF_THE_DAY all in the
same batch file with no plumbing, data conversion or additional
machination.

But if I could do this in MySQL + R, well in a few years I'd have the down
payment for an Audi TT instead of having given it to SAS Institute.

If someone found a practical way to make this work -- pls let me know.  And
thanks in advance.

Jack Williamson
LECG
2049 Century Park East, Suite 2300
Los Angeles CA 90067
323-683-5004
jwilliamson at lecg.com