I reproduce from memory my exhaustive look into this issue. RODBC uses the Microsoft ODBC DLL's developed by Microsoft. These DLL's perform an automatic determination of column type based on the contents of the first N rows of cells in each column, where N [0,16]. N may be set in the Windows system registry, and there are a few other things that may be set in the system registry which control how the DLL parses an Excel spreadsheet. Unfortunately, the Microsoft DLL's do not always pay attention to the registry settings and do not always interpret them in the same manner. The end result is that no matter what you do with RODBC, and no matter how the authors of RODBC re-write it, some Excel spreadsheets will always be unreadable via RODBC given particular insidious combinations of data in some columns of your spreadsheet. (until such time as Microsoft fixes their DLL bugs, I mean features) I have some faint recollection that the Microsoft DLL incorrectly parses a column with non-empty rows due to some formatting issue of those particular columns, which I was unable to cure by re-formatting the source worksheet. I have had to resort to using the gdata package which runs a Perl script "xls2csv.pl", which converts an Excel spreadsheet to CSV, for a few Excel spreadsheets which exhibit the particular anomalies preventing use of RODBC.
Leif Kirschenbaum Senior Yield Engineer Reflectivity, Inc. (408) 737-8100 x307 [EMAIL PROTECTED] > Message: 21 > Date: Mon, 9 Jan 2006 18:06:49 +0100 > From: "Fredrik Lundgren" <[EMAIL PROTECTED]> > Subject: Re: [R] "Missing value representation in Excel before > extraction to R with RODBC" > To: "Prof Brian Ripley" <[EMAIL PROTECTED]>, "Petr Pikal" > <[EMAIL PROTECTED]> > Cc: R-help <[email protected]> > Message-ID: <[EMAIL PROTECTED]> > Content-Type: text/plain; format=flowed; charset="iso-8859-1"; > reply-type=response > > Dear list, > > Well, those columns in Excel that starts with NA (actually 8 > NA's in my > case) is imported as all NA in R but if the columns starts > with at least > 3 cells with values (i.e not NA) the are imported correctly > to R. When > as.is=TRUE is used a simular conversion takes place but now > as all <NA> > and dates are represented as date-and-time. > Is there any way to get this correct even when the Excel > columns start > with several NA's? > > Sincerely > Fredrik > > > ----- Original Message ----- > From: "Prof Brian Ripley" <[EMAIL PROTECTED]> > To: "Petr Pikal" <[EMAIL PROTECTED]> > Cc: "Fredrik Lundgren" <[EMAIL PROTECTED]>; "R-help" > <[email protected]> > Sent: Monday, January 09, 2006 9:36 AM > Subject: Re: [R] "Missing value representation in Excel before > extraction to R with RODBC" > > > > On Mon, 9 Jan 2006, Petr Pikal wrote: > > > >> Hi > >> > >> I believe it has something to do with the column identification > >> decision. When R decides what is in a column it uses only > some values > >> from the beginning of a file. > > > > Not R, Excel. Excel tells ODBC what the column types are. > > ______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
