Forum: vegan & R 2.13.0
1) The output of cmdscale() changed in R 2.13.0. Function capscale() in vegan tries to sanitize the old output, but the new output is already sane, and unnecessary sanitization causes an error. The error may appear in capscale(..., add = TRUE), and because of this error, BiodiversityR package fails its test in example(multiconstrained). We released a new vegan version (1.17-10) to fix the problem, and you should upgrade to vegan_1.17-10 if you plan to use capscale() with argument add = TRUE, or use BiodiveristyR.
2) R 2.13.0 introduced a new generic function nobs() that returns the number of observations in a fitted model. We added nobs() functions for vegan model objects, but we did not include them in the latest release version, because we didn't realize that they would be in full use in R 2.13.0 already. Currently, some functions give unnecessary warnings. The issue was reported for ordiR2step() in vegan-help forum in R-Forge, but probably also ordistep() and step() are touched. This is not a severe problem and we won't release a new version only to fix this issue, but the fix will be included in the next release. If the issue is disturbing, you can either use the current development version (that had nobs() since Feb 24, 2011) or download the file nobs.R from vegan/pkg/vegan/R directory in R Forge. If the issue looks really bad, please contact us and we revise our policy and release a new version of vegan.
Jari Oksanen - 2014-09-18 15:58 -
Jari Oksanen - 2014-09-18 15:57 -
|Topic||Topic Starter||Replies||Last Post|
|Veagn input file||shashank gupta||0||2015-01-30 10:57|
|Vegan Package : input for Vegan from QIIME||shashank gupta||0||2015-01-21 06:08|