[R] How to import SAS data in R?
djnordlund at frontier.com
Fri Jul 6 23:03:21 CEST 2012
> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
> On Behalf Of C W
> Sent: Friday, July 06, 2012 12:30 PM
> To: David Winsemius
> Cc: r-help help
> Subject: Re: [R] How to import SAS data in R?
> Has anyone read SAS file .sas7bdat into R. The above suggestions don't
> On Fri, Jul 6, 2012 at 3:06 PM, David Winsemius
> <dwinsemius at comcast.net>wrote:
> > On Jul 6, 2012, at 2:49 PM, C W wrote:
> > Hey, David
> >> table(count.fields()) is telling me have 11 columns, but I have way
> >> more like 30 columns.
> >> > table(count.fields("**persistency.csv"))
> >> 1 2 3 4 5 6 7 8 9 11
> >> 439384 39617 16130 21993 12556 1900 988 713 61 1
> > I don't usually put back the list address when responding to a private
> > email, but I'm breaking my rule. You should not send private follow-ups
> > threads that start on the list unless there is something that other
> > might not benefit from seeing.
> > If it is a CSV file then you need to tell count.fields to use commas:
> > table(count.fields("**persistency.csv", sep=","))
> > --
> > David.
> >> -M
> >> On Fri, Jul 6, 2012 at 2:36 PM, David Winsemius
> <dwinsemius at comcast.net>
> >> wrote:
> >> On Jul 6, 2012, at 1:39 PM, C W wrote:
> >> Quick question, what the memory size in R?
> >> I converted to CSV, but only 53300 of the 1,000,000 rows were read in.
> >> Did
> >> R run out of memory? If so, is there a work around?
> >> You probably have mismatched quotes. Consider using quote="". Also
> >> consider doing this:
> >> table(count.fields(file-name)) # with a valid file name
> >> That count.fields function is very useful since it accepts the same
> >> arguments as the read.tables functions, with defaults of:
> >> quote = "\"'", skip = 0, blank.lines.skip = TRUE, comment.char = "#")
> >> --
> >> David.
> >> Thanks,
> >> Mike
> >> On Fri, Jul 6, 2012 at 1:24 PM, Duncan Murdoch
> <murdoch.duncan at gmail.com>
> >> **wrote:
> >> On 06/07/2012 1:11 PM, C W wrote:
> >> Hi all
> >> I have a large SAS data set, how do I get it read in R?
> >> The data is too big (about 400,000 rows by 100 columns) to be saved as
> >> Excel file. How should I get it read in R? Any packages? I don't
> >> to
> >> find any.
> >> You could write it out in some plain delimited format, e.g. CSV or
> >> tab-delimited. Watch out for special characters in strings that
> confuse R
> >> when it reads it in (e.g. commas in unquoted CSV strings, quotes within
> >> strings, etc.)
> >> Duncan Murdoch
> >> [[alternative HTML version deleted]]
> >> ______________________________**________________
> >> R-help at r-project.org mailing list
> >> https://stat.ethz.ch/mailman/**listinfo/r-
> >> PLEASE do read the posting guide http://www.R-project.org/**
> >> posting-guide.html <http://www.R-project.org/posting-guide.html>
> >> and provide commented, minimal, self-contained, reproducible code.
> >> David Winsemius, MD
> >> West Hartford, CT
> > David Winsemius, MD
> > West Hartford, CT
As has been pointed out there are many ways to get SAS data into R. Some are easier than others, depending on what the structure of the data is, and whether or not you have access to SAS. The sas7bdat package works fine for reading many (but not all) SAS datasets with the extension of .sas7bdat
Do you have access to SAS or do you just have the SAS dataset? Is this a one time thing, or are you going to be moving data back and forth between SAS and R a lot? I have found that the easiest method for me to move data back and forth between SAS and R is to use the Stata file format as a common format. Both SAS (v9.2 with SAS ACCESS to PC File Formats) and R read and write Stata format data files just fine. Again if you have SAS, the sas.get function in the Hmisc package is very useful.
If you don't have SAS, then you will need to follow the advice that has already been give (and follow it carefully).
Hope this is helpful,
Bothell, WA USA
More information about the R-help