[this message needed manual improvement by the mailing list administrator since it was `HTMLified' .. ``please do not''] Apologies for bothering you about a fairly trivial matter. I have been getting some inconsistencies with the display digits in R V1.5. I have been using the hypergeometric distribution function, and have found that when printing out the results from this function the display does not seem to correspond to the number of digits requested in the options settings. Maybe this indicates that the vector hasn't been generated in the proper way, or something, but I thought I'd report it all the same. [Previously saved workspace restored]> sqrt(2)[1] 1.414214> options(digits=4) > sqrt(2)[1] 1.414> options(digits=3) > for(j in 1:30){temp[j]<-sum(dhyper(0:6, j, 30 - j, 6)[3:7])} > temp[1] 0.0000 0.0345 0.0936 0.1691 0.2543 0.3438 0.4333 0.5195 0.6002 0.6736 [11] 0.7389 0.7956 0.8437 0.8835 0.9157 0.9410 0.9603 0.9744 0.9844 0.9912 [21] 0.9954 0.9979 0.9992 0.9998 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000> sqrt(2)[1] 1.41> temp[1] 0.0000 0.0345 0.0936 0.1691 0.2543 0.3438 0.4333 0.5195 0.6002 0.6736 [11] 0.7389 0.7956 0.8437 0.8835 0.9157 0.9410 0.9603 0.9744 0.9844 0.9912 [21] 0.9954 0.9979 0.9992 0.9998 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000> options(digits=4) > temp[1] 0.00000 0.03448 0.09360 0.16913 0.25435 0.34383 0.43330 0.51954 0.60018 [10] 0.67361 0.73889 0.79558 0.84368 0.88352 0.91571 0.94100 0.96026 0.97443 [19] 0.98444 0.99116 0.99540 0.99788 0.99917 0.99976 0.99996 1.00000 1.00000 [28] 1.00000 1.00000 1.00000> sqrt(2)[1] 1.414> x<-sqrt(2) > x[1] 1.414 While I'm writing, I'd like to thank all of you who are working on this R project. I think it is very valuable work. I have a friend who after working as a computer consultant for many years had to retire after having a stroke which left him unaffected mentally but without the use of his left hand. He now lives on an invalid pension, and has gone back to doing scientific work in his retirement. There is no way he could afford commercial software. R has given him the tools to do good quality graphics as required by journal editors without which he would not have got his papers published. There must be many people who have something to contribute but who work outside institutions and don't have the means to purchase commercial licences. And Third World countries should benefit greatly. Regards, Geoff Gordon -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On Tue, 6 Aug 2002 gordong@epa.nsw.gov.au wrote:> Apologies for bothering you about a fairly trivial matter. I have been > getting some inconsistencies with the display digits in R V1.5. I have been > using the hypergeometric distribution function, and have found that when > printing out the results from this function the display does not seem to > correspond to the number of digits requested in the options settings. Maybe > this indicates that the vector hasn't been generated in the proper way, or > something, but I thought I'd report it all the same.Could you point out where it does not give the number of digits requested? When you print out a vector R uses the same number of decimal places for all entries, so one entry will have the number of significant digits requested, but others may have more. BTW, your example does not create temp, and you haven't told us the platorm (as created by bug.report).> > [Previously saved workspace restored] > > > sqrt(2) > [1] 1.414214 > > options(digits=4) > > sqrt(2) > [1] 1.414 > > options(digits=3) > > for(j in 1:30){temp[j]<-sum(dhyper(0:6, j, 30 - j, 6)[3:7])} > > temp > [1] 0.0000 0.0345 0.0936 0.1691 0.2543 0.3438 0.4333 0.5195 0.6002 0.6736 > [11] 0.7389 0.7956 0.8437 0.8835 0.9157 0.9410 0.9603 0.9744 0.9844 0.9912 > [21] 0.9954 0.9979 0.9992 0.9998 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 > > sqrt(2) > [1] 1.41 > > temp > [1] 0.0000 0.0345 0.0936 0.1691 0.2543 0.3438 0.4333 0.5195 0.6002 0.6736 > [11] 0.7389 0.7956 0.8437 0.8835 0.9157 0.9410 0.9603 0.9744 0.9844 0.9912 > [21] 0.9954 0.9979 0.9992 0.9998 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 > > options(digits=4) > > temp > [1] 0.00000 0.03448 0.09360 0.16913 0.25435 0.34383 0.43330 0.51954 0.60018 > [10] 0.67361 0.73889 0.79558 0.84368 0.88352 0.91571 0.94100 0.96026 0.97443 > [19] 0.98444 0.99116 0.99540 0.99788 0.99917 0.99976 0.99996 1.00000 1.00000 > [28] 1.00000 1.00000 1.00000 > > sqrt(2) > [1] 1.414 > > x<-sqrt(2) > > x > [1] 1.414-- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
gordong@epa.nsw.gov.au wrote:> > [this message needed manual improvement by the mailing list > administrator since it was `HTMLified' .. ``please do not''] > > Apologies for bothering you about a fairly trivial matter. I have been > getting some inconsistencies with the display digits in R V1.5. I have been > using the hypergeometric distribution function, and have found that when > printing out the results from this function the display does not seem to > correspond to the number of digits requested in the options settings. Maybe > this indicates that the vector hasn't been generated in the proper way, or > something, but I thought I'd report it all the same. > > [Previously saved workspace restored] > > > sqrt(2) > [1] 1.414214 > > options(digits=4) > > sqrt(2) > [1] 1.414 > > options(digits=3) > > for(j in 1:30){temp[j]<-sum(dhyper(0:6, j, 30 - j, 6)[3:7])} > > temp > [1] 0.0000 0.0345 0.0936 0.1691 0.2543 0.3438 0.4333 0.5195 0.6002 0.6736 > [11] 0.7389 0.7956 0.8437 0.8835 0.9157 0.9410 0.9603 0.9744 0.9844 0.9912 > [21] 0.9954 0.9979 0.9992 0.9998 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 > > sqrt(2) > [1] 1.41 > > temp > [1] 0.0000 0.0345 0.0936 0.1691 0.2543 0.3438 0.4333 0.5195 0.6002 0.6736 > [11] 0.7389 0.7956 0.8437 0.8835 0.9157 0.9410 0.9603 0.9744 0.9844 0.9912 > [21] 0.9954 0.9979 0.9992 0.9998 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 > > options(digits=4) > > temp > [1] 0.00000 0.03448 0.09360 0.16913 0.25435 0.34383 0.43330 0.51954 0.60018 > [10] 0.67361 0.73889 0.79558 0.84368 0.88352 0.91571 0.94100 0.96026 0.97443 > [19] 0.98444 0.99116 0.99540 0.99788 0.99917 0.99976 0.99996 1.00000 1.00000 > [28] 1.00000 1.00000 1.00000 > > sqrt(2) > [1] 1.414 > > x<-sqrt(2) > > x > [1] 1.414That's not a bug. R prints the number of significant digits, as expected. For the second element in your vector temp, there are four sifnificant digits printed: the "3448" within 0.03448. Just try some examples: 0.123456789 # [1] 0.1234568, seven significant digits options(digits=4) 0.123456789 # [1] 0.1235, as expected 0.001234567 # [1] 0.001235, still 4 significant digits So, I cannot see any bug. Uwe Ligges> While I'm writing, I'd like to thank all of you who are working on this R > project. I think it is very valuable work. I have a friend who after working > as a computer consultant for many years had to retire after having a stroke > which left him unaffected mentally but without the use of his left hand. He > now lives on an invalid pension, and has gone back to doing scientific work > in his retirement. There is no way he could afford commercial software. R > has given him the tools to do good quality graphics as required by journal > editors without which he would not have got his papers published. There must > be many people who have something to contribute but who work outside > institutions and don't have the means to purchase commercial licences. And > Third World countries should benefit greatly.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._