2013-03-19 Søren Højsgaard * parseGroupFormula function added. The function will take a "group formula" with a vertical bar apart into components. * createFunBy function added. The function can create groupwise functions. * documentation of lmBy improved * various utility functions related to lmBy added. * Version 4.5-6 uploaded 2012-11-16 Søren Højsgaard * Rmarkup removed; the markdown package does a better job. * Version 4.5-5 uploaded 2012-05-28 Søren Højsgaard * Updates of Rmarkup * Version 4.5-4 uploaded 2012-05-19 Søren Højsgaard * Updates of Rmarkup * Version 4.5-3 uploaded 2012-02-17 Søren Højsgaard * descStat function added. Thanks to Gregor Gorjanc. * popMeans extended so that the grid values appear in the output by default * Version 4.5.2 uploaded 2012-01-29 Søren Højsgaard * Minor enhancement to Rmarkup() * Version 4.5.1 uploaded 2012-01-18 Søren Højsgaard * popMeans method for lme objects added * lsmeans aliases for popMeans removed * Version 4.5.0 uploaded 2012-01-18 Søren Højsgaard * Small changes to package structure * Version 4.4.6 uploaded 2012-01-17 Søren Højsgaard * URL updated * Rscript2HTML changed to Rmarkup * Version 4.4.5 uploaded 2011-12-30 Søren Højsgaard * Change of email address * Version 4.4.4 uploaded. 2011-11-01 Søren Højsgaard * Function lmBy (and related methods) added * Version 4.4.3 uploaded. 2011-11-01 Søren Højsgaard * Dataset milkman added * Version 4.4.2 uploaded. 2011-10-23 Søren Højsgaard * funBy function added (documentation to be done) * scaleBy function added * Minor bug fixed in summaryBy * popMeans extended to mer objects * KRmodcomp, PBmodcomp and friends moved to the pbkrtest package * Version 4.4.1 uploaded. 2011-06-05 Søren Højsgaard * subSeq function speeded up by considerably * KRmodcomp: Kenward-Rogers method for calculating denominator degrees of freedom for F--tests in mixed models (as fitted with lmer) * PBmodcomp, BCmodcomp, PBrefdist: Functions for calculating p-values based on parametric bootstrap for tests in mixed models (as fitted with lmer) * Version 4.4.0 uploaded 2011-05-05 Søren Højsgaard * HTMLreport changed to Rscript2HTML * timeSinceEvent function enhanced * Version 4.3.1 uploaded. 2011-03-28 Søren Højsgaard * popMeans (lsMeans) and linMeans functions added. * A vignette describing the use of popMeans added. * Functions popMatrix and linMatrix added * Version 4.3.0 uploaded 2011-02-02 Søren Højsgaard * Bug fixed in timeSinceEvent * More work done on HTMLreport * Version 4.2.3 uploaded 2011-01-12 Søren Højsgaard * HTMLreport function extended * Version 4.2.1 uploaded 2010-12-12 Søren Højsgaard * HTMLreport function has been added. * Version 4.2.0 has been uploaded. 2010-11-24 Søren Højsgaard * timeSinceEvent function has been added * summaryBy now takes the additional argument full.dimension which when TRUE causes the result to have the same number of rows as the input dataframe. * Version 4.1.2 has been uploaded. 2010-11-11 Søren Højsgaard * recodevar deprecated. * recodeVar introduced with extended functionality. * subSeq function added. * renameCol function added. * Version 4.1.1 has been uploaded. 2010-11-09 Søren Højsgaard * lsmeans function has been added. * Version 4.1.0 has been uploaded. 2010-05-05 Søren Højsgaard * The sugar beets dataset 'beets' has been added * doBy no longer depends on Hmisc * Version 4.0.6 has been uploaded 2009-10-21 Søren Højsgaard * orderBy applied to a dataframe with one column only returned a vector. Has been fixed so that a dataframe is returned. * Version 4.0.5 uploaded 2009-10-17 Søren Højsgaard * Bug in NAMESPACE file fixed * Version 4.0.4 uploaded 2009-10-17 Søren Højsgaard * esticon method for 'mer' objects added. (Was previously for 'lmer' objects). * Version 4.0.3 uploaded 2009-09-07 Søren Højsgaard * Bug in summaryBy fixed: Strata did not match their values (as defined by the rh.variables). Thanks to Pascal Hirsch for pointing this out. * Version 4.0.2 uploaded 2009-08-17 Søren Højsgaard * Bug in summaryBy fixed: id vars were included in the stratification of data which they should not be. Thanks to Joshua Rest for pointing this out. * Version 4.0.1 uploaded. 2009-07-28 Søren Højsgaard * coxph method added for the esticon function. Thanks to Alessandro A Leidi. * splitBy changed so that names of the entries in the list match the levels of the factors (it used to be that the names matched the levels of the factors in reverse order). * The formula in splitBy can now also be a character vector. * which.maxn and which.minn functions added * Version 4.0.0 uploaded 2009-01-12 Søren Højsgaard * splitBy had inconsistent output across cases. Fixed now * Version 3.8 uploaded 2008-12-21 Søren Højsgaard * splitBy failed if by-variables were constant. Fixed now. * Version 3.7 uploaded. 2008-10-14 Søren Højsgaard * splitBy now follows convention that first variable varies fastest. * Version 3.6 uploaded 2008-10-14 Søren Højsgaard * Updated description of dietox data * Version 3.5 uploaded 2008-09-26 Søren Højsgaard * Small bug in summaryBy fixed * Version 3.4 uploaded 2008-09-26 Søren Højsgaard * Small bug in summaryBy fixed * Version 3.3 uploaded 2008-09-18 Søren Højsgaard * Longer formulas now accepted (bug caused by a naive use of deparse()) * Version 3.2 uploaded 2008-06-13 Søren Højsgaard * Bugs in variable names in summaryBy fixed * Small update of documentation * Version 3.1 uploaded 2008-05-20 Søren Højsgaard * summaryBy speeded up; postfix argument removed * Version 3.0 uploaded 2008-03-26 Søren Højsgaard * Minor bugs removed. * Version 2.3 uploaded 2008-02-13 Søren Højsgaard * firstobs and lastobs functions added * Version 2.2 uploaded 2007-12-05 Søren Højsgaard * splitBy: Results becomes a splitByData object. Two attributes are added: 1) grps, which is the grouping factors coded into a single variable. 2) idxvec, (a list) which holds the position of each row in the original data set. * lapplyBy: Following the changes above, the results of lapplyBy follows the ordering given by formula. * subsetBy: Now subset is a logical expression, not a character string * Version 2.1 uploaded 2007-09-30 Søren Højsgaard * DESCRIPTION file change from Depends: Hmisc to Imports: Hmisc; * importFrom(Hmisc,mApply) added so that not all of Hmisc is loaded. 2007-05-01 Søren Højsgaard * summaryBy: Behaviour with . on rhs of formula when data contains no factors has been corrected. * Version 1.9 uploaded 2007-05-01 Søren Højsgaard * Miscellaneous corrections in the .Rd files * lapplyBy function has been added * Vignette has been updated * Version 1.8 uploaded 2007-04-21 Søren Højsgaard * orderBy changed so that signs in formula determines whether to sort ascendingly or decreasingly 2007-03-21 Søren Højsgaard * esticon extended to handle gls objects. 2006-12-14 Søren Højsgaard * Version 1.6 uploaded * .asNumericMatrix2 etc now work correctly with date formats 2006-11-05 Søren Højsgaard * Version 1.5 uploaded * Typo in package title corrected 2006-11-05 Søren Højsgaard * Version 1.4 built and uploaded * A warning from Kurt H (regarding gsub...) has been fixed * Internal functions have been 'dotted' * A NAMESPACE file has been added * A file describing how to generate the NAMESPACE file automatically has been added. 2006-10-23 Søren Højsgaard * Version 1.3 built and uploaded * Bug in splitBy corrected so that it works when rhs includes a character variable. 2006-10-23 Søren Højsgaard * Version 1.2 built and uploaded * Description of codstom data updated. * matrix2dataFrame2 and subsAttr2 added * postfix argument in summaryBy can be a list 2006-10-17 Søren Højsgaard * Version 1.1 built and uploaded * codstom.Rd updated * asNumericMatrix2 added because asNumericMatrix in Hmisc does not work if a variable is of type 'character' 2006-10-16 Søren Højsgaard * Version 1.0 built and uploaded * summaryBy: More issues about naming of variables have been sorted out * summaryBy: Argument order=TRUE has been added * summaryBy: prefix=NULL changed to postfix=NULL, argument p2d has been added 2006-10-10 Søren Højsgaard * Version 0.9 built and uploaded * summaryBy assigns better names to generated variables * A vignette has been created 2006-10-03 Søren Højsgaard * Version 0.8 built and uploaded * Handles now ordered factors * Fixed bug so that summaryBy now works with do.call 2006-10-01 Søren Højsgaard * Version 0.7 built and uploaded * splitBy is now based on functions in the Hmisc package for representing dataframes as matrices, which makes the function work also for 'large' (10000 rows) datasets. * summaryBy can now take . on both lhs and rhs of "~" 2006-09-27 Søren Højsgaard * Various bugs have been fixed 2006-09-18 Søren Højsgaard * summaryBy has been reimplemented. * summaryBy takes . as argument on both lhs and rhs in formula * summaryBy, splitBy takes drop=FALSE argument (just like split does) * transformBy function has been added 2006-02-05 Søren Højsgaard * sampleBy modified allow for also 'systematic samples', e.g. every 5th row. * summaryBy takes ... arguments to FUN. No default NA action is taken. 2006-02-02 Søren Højsgaard * A sampleBy function has been added. * A subsetBy function has been added. 2006-01-26 Søren Højsgaard * Function summaryBy modified so one can write log(x) on the left hand side of '~' * Function summaryBy modified such that NAs are removed. What to do about Inf?? 2006-01-17 Søren Højsgaard * Function summaryBy modified to take . as lhs in formula and to take idvar = ~formula * Functions powerBy, histBy, qqnormBy have been removed (figure out how to do histBy and qqnormBy using xyplot() in lattice).