[R] read.spss warning: unrecognized record type??

Peter Dalgaard p.dalgaard at biostat.ku.dk
Fri Aug 27 09:17:45 CEST 2004


"Michael Jerosch-Herold" <jeroschh at ohsu.edu> writes:

> When using read.spss (library: 'foreign') I get the following warning
> message:
> 
> Warning message: 
> E:/R4win/mesamri.sav: Unrecognized record type 7, subtype 13
> encountered in system file. 
> 
> I don't see anything wrong with record #7 in the database I am trying
> to read in, but I suspect that the warning message does not refer to a
> specific record, but a "variable" type. Is this correct? And what does
> "subtype 13" mean?

Well foreign doesn't know either, and that's the problem... Most
likely you have one of your SPSS variables coded in an unusual manner
(do they all come though alright?), or it is something else in the
meta-info that foreign cannot interpret.
 
> I basically have a data table with records (cases) in rows, and various
> variables for each record/case in the columns. Again, I suspect that I
> am using "record" in a different sense then meant by the above warning
> message.

Yes. System files are typically organised in "records" which are not
the actual data records but pieces of meta-information. As in (I'm
just making this up, read the sources for the exact definitions)

type=1 subtype=1 length=22 data1="My very important data"

which is compactly stored as (<..> meaning "binary coding of")

<1><1><22>My very...

so that on reading, the program sees the first two bytes and then
knows that this is a "main header" and that it will be followed by a
2-byte length and then as many characters as the length indicates.

There are other record types for individual variable names, variable
types, variable labels, value labels for categorical data, etc...

 
> Based on this warning, is "record type 7" discarded when the data are
> read in?

Possibly, but you have the data to check and we don't....
 
> Thank you in advance for shedding some light on this!

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907




More information about the R-help mailing list