when I start the GUI version of Splus on a Sun Solaris machine I get the following warning: SUN-V880% Warning: Cannot convert string " -b&h-lucidasans-medium-r-normal-sans-*-140-*-*-p-*-iso8859-1" t
Hello S-users, thank you very much to Joseph Lucke and to Marc Schwartz for their replies. I think I didn't voice clearly enough what I was trying to do. I would like to calculate Cramer's V statisti
Hello S-users, I am trying to use barchart to plot the data from a small data.frame. No matter how the rows of the input data.frame are ordered barchart plots them always in the same order; I would l
Thanks to all who responded, Douglas Bates, James Holtman, John Payne, Bill Dunlap, Giorgi Roch and Jim Pratt. Two solutions were suggested. 1) Convert the variable "v" to a factor, or ordered factor
Hello S-users, anyone aware of an implementation to calculate the Cramer's V statistics? Thanks, Gérald Jean Analyste-conseil (statistiques), Actuariat télephone : (418) 835-4900 poste (7639) télec
Hello S-users, I have a user function which returns an object and assigns attributes to this object through the builtin function "structure". I have used my function for years and it works well. Unde
Hello S-users, I am trying to build a glm model using Gordon Smyth's "tariff" function. The function was specifically written to build double glm's (modeling mean and dispersion simultaneously) for i
Hello S-users, I am running Splus6.1, 64 bit version for Unix (Solaris). I am currently trying to read into S+ a large data set, over 4 000 000 records by 58 variables. It took barely over 15 minutes
Last week I posted a request regarding the reading of large data sets. I had one or two replies suggesting the preprocessing of the data using Perl or similar languages, unfortunately I don't know P
Hello S-users, when creating factors is it possible to merge levels of a character variable as the factor is created? In other words can the following be done in one operation. ttt <- factor(ttt.char
Hello S-users, I have a glm model. I wanted to have linear predictors, not centered as predict does, for only some of the terms; I started to play with the linear predictors and matrix multiplication
I am using a 64-bit version of Splus 6.0 on Sun-Solaris. I am trying to import a large SAS data set which is in transport format. I am getting the following error. + '/home/jeg002/splus/AutoROC/AB.x
Hello S-users, we recently installed a 64-bit version of Splus6 for Unix. Not sure if I am blind or what, but I can't find F. Harrell's librairies Hmisc and Design in the Splus distribution; for that
I am currently testing a 64-bit version of Splus6.0 on a Sun Solaris machine. I am wondering why memory is not released after a simple subsetting operation? Here is a comparison between an NT and a
Hello S-users, I got a single reply, so far, to my querries about S+ memory usage under Solaris. The reply is from David Smith and, as usual, it is very informative, thanks David. It turns out that m
Hello S users, under Windows NT we can address 2GB of memory, what would be the limit under Linux? In the very near future we will probably install Linux on my machine and get a Linux version of S+6,
Hello S-users, in the very near future I will be testing a 64-bit version of S+6 on Solaris. In order to conduct the tests I would need two user written librairies that I am currently using in S+6 fo
Hello S-users, I am trying to rotate axis labels using the "postscript" device, they are too long to fit if not rotated. I have no problems doing that using the "graphsheet" device, but to do so one