Changes in version 0.8.10 (2025-06-19) o fix outstanding R CMD check problems Changes in version 0.8.9 (2022-04-12) o fix outstanding R CMD check problems Changes in version 0.8.8 (2018-10-23) o fix outstanding R CMD check problems Changes in version 0.8.7 (2017-08-06) o fix outstanding R CMD check problems Changes in version 0.8.6 (2016-10-21) o fix outstanding R CMD check problems Changes in version 0.8.5 (2014-04-23) o fix outstanding R CMD check problems Changes in version 0.8.4 (2011-01-28) o fix spelling mistake (indicies -> indices), thanks to Stavros Macrakis Changes in version 0.8.3 (2009-04-28) o better rename example o When removing missing values in melt, look only at measured variables, not id variables o Fixes to documentation bugs revealed by new parser Changes in version 0.8.2 (2008-11-04) o fixed bug where missing fill values where not getting correctly filled o fill value defaults to fun.aggregate applied to zero-length vector. This produces better values in a wide variety of situations, for example missings will be filled with 0's when length or sum is used. This may require setting fill = NA for aggregation functions that previously return NA, like sd and var. Changes in version 0.8.1 (2008-09-24) Melt o character dimension names are processed by type.convert o by default, treat character and factor variables as id variables (i.e. integer variables no longer default to being id vars) o ... now passed on to melt in melt.list (thanks to Charles Naylor) Cast o missing values in subsetting are now correctly dropped to match behaviour of subset() o tweaks to cast and recast to make it easier to pass in formulas created elsewhere o allow user to specify column used for values, guessing if necessary, and giving a warning message if value column does not exist o improve error messages when melt or casting parameters incorrectly specified General o now depends on the new plyr package for many of the workhorse functions