[R] avoiding excel's odbc limit on number of columns

[Alejandro Munoz del Rio]

Dear R-Helpers,

I would like to read an Excel .xls file via RODBC. I have successfully run
the example in p. 18 of "R Data Import/Export". The problem I am facing is
that Excel's ODBC driver seems to have a limit on the number of
fields/columns (output below). I haven't found any documentation on what
this limit L might be, but I know that 128 <= L < 256.

Does anyone know of a way to avoid "L" via
- the arguments to sqlQuery() or sqlGetResults()?
- an SQL select statement that can subset the columns/fields and index a
range of columns?
- otherwise?

[system info: win98 with 128Mb RAM, R v1.6.1, RODBC v. 1.0-1, excel 2000,
odbc driver v. 03.51; my knowledge of sql = $\epsilon$.]

Gratefully,

alejandro

> #Sheet1 has 92 rows and 256 columns; its structure is roughly as follows:
> # [blank]  0  1  2 ...  254
> #         0 70 70 71 ...  63
> # [89 rows deleted]
> #       90 57 56 52 ...  37
> frame1 <- sqlQuery(channel, "select * from [Sheet1$]")
> frame1
[1] "[RODBC] ERROR: Could not SQLExecute"
[2] "S1001 -1040 [Microsoft][ODBC Excel Driver] Too many fields defined."
> odbcGetInfo(channel)
[1] "EXCEL version 08.00.0000. Driver ODBC version 03.51"
> version
         _
platform i386-pc-mingw32
arch     i386
os       mingw32
system   i386, mingw32
status
major    1
minor    6.1
year     2002
month    11
day      01
language R