On 23 Apr 2002, Peter Dalgaard BSA wrote:
richard_raubertas@merck.com writes:
From: Huntsinger, Reid Sent: Monday, April 22, 2002 5:15 PM To: Raubertas, Richard Subject: RE: R problem [ ... ] This happens because of the attempt to convert. The function type.convert is called on columns with unspecified types; this tries to convert to "logical, integer, numeric or complex" in that order (first one that is successful wins). You have some ids like 21428-2 which type.convert doesn't like as numeric but will treat as complex. This in turn is due to R's hand-rolled "strtoc" function, which returns 0+0i for something it doesn't like. This is a valid value, so the conversion "works", and all values are converted this way... You or I should probably alert the R people. That's not very good behavior.
Right, but not something we can fix for 1.5.0. Another version of the same effect is
scan(what=0+0i)
1: 21419-2 2: Read 1 items [1] 0+0i A workaround would be to use the colClasses argument, or as.is.
Actually it is already fixed: it was a very simple fix. The problem was being detected: it was just not being reported. As this was just a case of what is done in an error condition (it can't affect any valid input) I thought it safe to fix a week before release.
Brian D. Ripley, ripley@stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272860 (secr) Oxford OX1 3TG, UK Fax: +44 1865 272595 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._