NEWS in package ‘reshape’

---

Changes in version 0.8.9

• fix outstanding R CMD check problems

Changes in version 0.8.8

• fix outstanding R CMD check problems

Changes in version 0.8.7

• fix outstanding R CMD check problems

Changes in version 0.8.6

• fix outstanding R CMD check problems

Changes in version 0.8.5

• fix outstanding R CMD check problems

Changes in version 0.8.4

• fix spelling mistake (indicies -> indices), thanks to Stavros Macrakis

Changes in version 0.8.3 (2009-04-27)

• better rename example
• When removing missing values in melt, look only at measured variables, not id variables
• Fixes to documentation bugs revealed by new parser

Changes in version 0.8.2 (2008-11-04)

• fixed bug where missing fill values where not getting correctly filled
• 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 (2008-05-01)

Melt

• character dimension names are processed by type.convert
• by default, treat character and factor variables as id variables (i.e. integer variables no longer default to being id vars)
• ... now passed on to melt in melt.list (thanks to Charles Naylor)

Cast

• missing values in subsetting are now correctly dropped to match behaviour of subset()
• tweaks to cast and recast to make it easier to pass in formulas created elsewhere
• allow user to specify column used for values, guessing if necessary, and giving a warning message if value column does not exist
• improve error messages when melt or casting parameters incorrectly specified

General

• now depends on the new plyr package for many of the workhorse functions