On 19 Apr 2008, at 12:01 AM, Peter Dalgaard wrote: > Antony Unwin wrote: >> >>> A couple other maybe not all that trivial things to do is to >>> improve the data import (it is losing out on most of the things >>> that I tried) >> >> Now what would Brian say to a comment like that? Please insert >> your favourite put-down here: >> >> **** **** > He'd usually say something about reproducible examples, just like > you... > > Don't get me wrong. I realize that the Open dialog in DataTable is > designed to read one, simple data format, not all of them. I just > couldn't easily figure out which one that was. > >> And then perhaps you would be kind enough to let us know in a >> little more detail what hasn't worked for you. > Anything in the data directory of the ISwR package, for instance > (whitespace separated and comma separated files, mostly). I suspect > that the format it _will_ read is TAB separated, but it would be > nice if it said so somewhere. (Curiously, thuesen.txt does have > tabs, but not equally many in the header as in the data lines, so it > can't be read at all, the others just get all fields and headers > jammed into one.)
Thanks for making this clear. The way to load datafiles is to use the first command in the "File" menu "Load Datafile". This brings up a dialog box with the default options set (TAB separated, as you suspected) and with the opportunity to change them to whatever you want. Header lines define how many variables you have and so must have the same number of tabs as in the data lines to read the file successfully. Once you missed "Load Datafile" and went hunting for other ways of loading data, you were doomed to have problems. The "Data Table" command and its "Open" command are placeholders for possible future development and should be greyed out or left out rather than made available in their current state. Sorry. No one else has reported this problem and I hope no one else has had it. There are two short pdf files on JGR's webpage providing documentation and the Statistical Computing and Graphics Newsletter one has the information you needed. Otherwise writing to me or to Markus (Helbig) or Martin (Theus) or Simon (Urbanek) would have got a response. Writing to R-help is maybe a bit of overkill, but I'm glad you raised the issue. This is a lesson for us all that we never seem to learn: it's worth looking at the documentation, perhaps especially with R packages as everyone thinks a little bit differently and so does things a little bit differently. On the subject of whether to use manuals or not and looking for help, do have a look at Eddie Izzard's routine (enter "Eddie Izzard computer" on Youtube). If Eddie Izzard used statistical software, he'd use R. Best regards Antony ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
