[R] RODBC

[Prof Brian D Ripley]

On Wed, 15 Nov 2000, Emmanuel Charpentier wrote:

> I saw in the R-digest a thread about RODBC.
> 
> 1) I think that this package is *extremely* valuable : it fulfills a need
> quite important in the case of iterative (e. g. periodic) analyses. No need to
> periodically return to your data and re-dumping them ...
> 
> 2) The current version has a serious bug if your datasets has missing numeric
> values. M. Lapsley, the original author, is aware of it, but has had no time
> yet to look at it ...

I have tried in the last two days to reproduce it, and I just can't.
I've tried RODBC to MySQL on Windows, and RODBC to PostgreSQL on Unix.
Every time I get blank strings mapped to NAs, in numeric (right) or
character (wrong) fields, and I have a fix for the latter.

So can you tell me (and Michael L) exactly what you did?

> I looked into the source to no avail. I'll have to dive in ODBC library
> mysteries (which are pure Swahili to me ...), and did not yet had time for
> this.
> 
> <OffTheTopOfMyHeadIDontHaveTheFilesOnHand>
> In interim, you can uncomment the old (much less efficient) of the odbcQuery
> routine, which does *not* exhibit this bug. Data are correct, speed seriously
> sucks ...
> </OffTheTopOfMyHeadIDontHaveTheFilesOnHand>
> 
> 3) If you happen to use PostgreSQL, the RPgSql package (also on CRAN) is much
> more interesting : it allows direct access to PostgreSQL classes (tables) as R
> objects, ensuring that any R access is done to up-to-date data.
> 
> Of course, OTOH, if you happen to trash your R data-access-objects, your
> original data are gone for good. Caveat emptor ...
> 
> 4) IMNSHO, the future lies somewhere in the (hypothetical) fusion of these two
> package : direct access to ODBC objects would be perfect (and might seriously
> alleviate problems due to the R static memory allocation model, which is an
> EPITA ...).

What R static memory allocation model?  Take a look at R-devel....
Notwithstanding that, I agree and am beginning to use databases quite a
bit.  Of course, one then sees how to improve the interfaces ....

One thing R-1.2.0 will come with is a Data Import/Export manual covering
all the database access packages (and much else).  A draft is available
on the R developer site (http://developer.r-project.org/R-data.pdf)
if you want to take a look.  (Comments to me please, and it is not
completely finished yet, and that includes not completely checked yet.)

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._