I know *what* happened (Calc reformatted the data in ways I did not want or expect). It is not end-of-line conventions; they reformatted the data leaving the structure intact. As to why/how, that could depend on the sequence of operations, so I thought to ask here to see if you had collectively either found something specific to do or to avoid.
Gnumeric is now freshly installed and will get some testing; if I don't care for it, I'll look more at emacs. I don't ask much of a spreadsheet (show/edit a grid and maybe hide/show columns for complex data sets), but it would be nice if it did not reformat everything every time I open a file :( So far, gnumeric successfully opened a file; I will be a little less trusting when it comes to saving one. Thanks!! Bill ________________________________________ From: Mike Marchywka [marchy...@hotmail.com] Sent: Wednesday, October 13, 2010 3:10 PM To: dwinsem...@comcast.net; Schwab,Wilhelm K Cc: r-help@r-project.org Subject: RE: [R] [OT] (slightly) - OpenOffice Calc and text files ---------------------------------------- > From: dwinsem...@comcast.net > To: bsch...@anest.ufl.edu > Date: Wed, 13 Oct 2010 14:52:21 -0400 > CC: r-help@r-project.org > Subject: Re: [R] [OT] (slightly) - OpenOffice Calc and text files > > > On Oct 13, 2010, at 1:13 PM, Schwab,Wilhelm K wrote: > > > Hello all, > > > > I had a very strange looking problem that turned out to be due to > > unexpected (by me at least) format changes to one of my data files. > > We have a small lab study in which each run is represented by a row > > in a tab-delimited file; each row identifies a repetition of the > > experiment and associates it with some subjective measurements and > > times from our notes that get used to index another file with lots > > of automatically collected data. In short, nothing shocking. > > > > In a moment of weakness, I opened the file using (I think it's > > version 3.2) of OpenOffice Calc to edit something that I had mangled > > when I first entered it, saved it (apparently the mistake), and > > reran my analysis code. The results were goofy, and the problem was > > in my code that runs before R ever sees the data. That code was > > confused by things that I would like to ensure don't happen again, > > and I suspect that some of you might have thoughts on it. > > > > The problems specifically: > > > > (1) OO seems to be a little stingy about producing tab-delimited > > text; there is stuff online about using the csv and editing the > > filter and folks (presumably like us) saying that it deserves to be > > a separate option. > > You have been little stingy yourself about describing what you did. I > see no specifics about the actual data used as input nor the specific > operations. I just opened an OO.o Calc workbook and dropped a > character vector, "1969-12-31 23:59:50" copied from help(POSIXct) into > > Have any of you found a nice (or at least predictable) way to use OO > > Calc to edit files like this? > > I didn't do anything I thought was out of the ordinary and so cannot > reproduce your problem. (This was on a Mac, but OO.o is probably going > to behave the same across *NIX cultures.) > > -- > David > > > If it insists on thinking for me, I wish it would think in 24 hour > > time and 4 digit years :) > > Is it possible that you have not done enough thinking for _it_? > > > I work on Linux, so Excel is off the table, but another spreadsheet > > or text editor would be a viable option, as would configuration > > changes to Calc. > > > > Bill Probably instead of guessing and seeing how various things react, you could go get a utility like octal dump or open in an editor that has a hex mode and see what happened. This could be anything- crlf convention, someone turned it to unicode, etc. On linux or cygwin I think you have "od" available. Then of course, if you know what R likes, you can use sed to fix it... ______________________________________________ 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.
