[Rd] Probable bug in RODBC 0.7.x (PR#700)

[emmanuel.charpentier@sap.ap-hop-paris.fr]

Full_Name: Emmanuel Charpentier
Version: R 1.1.1
OS: Windows 98 FE / Linux (Debian 2.2)
Submission from: (NULL) (164.2.255.244)


This bug report is pro forma. I've already reported it informally on R-help and
R-devel.

The current (0.7) version of RODBC has a problem with missing values. It
correctly
gets missing "character" values (as blanks), but puts the last non-missing value
of
the same variable in place of missing "numeric" values. Doubleplus ungood ...

Example :

In database : ==>     What I got in R    What I expected

1  Foo 1.0            1 "Foo" 1.0        1 "Foo" 1.0
2      2.0            2 ""    2.0        2 ""    2.0
3  Bar                3 "Bar" 2.0        3 "Bar" NA
4  Gee 4.0            4 "Gee" 4.0        4 "Gee" 4.0

This has been checked on Windows (links to remote Postgres, local MS-Access and
local
Excel) and Linux (links to Postgres, both local and remote, ODBC
manager=unixODBC,
drivers=unixODBC-bundled drivers).

A cursory preliminary look at the code was unsuccessful. I have to delve into
the
ODBC interface specification and learn how to debug C under R in order to go
further.

Michael Lapsley, author of the RODBC package, is aware of the problem, but has
no
fix yet (as of beginning of October).


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._