NEWS

Changes in 3.0.0 (2023-10-09)
-----------------------------

Other changes:

- Removed installXLSXsupport(), read.xls(), sheetCount(), sheetNames(),
  xls2csv(), xls2tab(), xls2tsv(), xls2sep(), and xlsFormats(). See gdata 2.19.0
  NEWS entry announcing their deprecation.

- Added aliases n_obs(), object_size(), and starts_with(). Deprecated
  gdata::object.size().




Changes in 2.19.0 (2023-05-06)
------------------------------

New functions:

- Added update.data.frame() to replace rows of a data frame.

New features:

- Added argument 'add.col.nums' to left() and right().

Other changes:

- Improved ll() so it handles S4 objects. Also accepts unit="GB" (contributed by
  Michael Chirico) and unit="B", now case-insensitive.

- Improved right() so it returns the rightmost columns of a matrix or data
  frame.

- Renamed first two arguments of startsWith() to 'x' and 'prefix'.

- Removed elem() - was deprecated in gregmisc/gdata 2.0.3 (2005-02-25).

- Excel file support is now deprecated, to be removed in the near future. At the
  user level, the functions installXLSXsupport(), read.xls(), sheetCount(),
  sheetNames(), xls2csv(), xls2sep(), xls2tab(), xls2tsv(), and xlsFormats() now
  raise a warning about the functionality being deprecated. Excel file support
  was first introduced in gregmisc/gdata 1.11.0 back in 2004, but today we have
  packages such as openxlsx, readxl, XLConnect, and xlsx offering dedicated
  Excel file support with more features. When Excel file support will be removed
  from the gdata package, it will result in lighter package maintenance and no
  more Perl warnings on the user side.

- Package maintainer is Arni Magnusson.




Changes in 2.18.0.1 (2022-05-10)
--------------------------------

Other changes:

- Syntax improvements to pass CRAN checks, packaged by Brian Ripley.




Changes in 2.18.0 (2017-06-05)
------------------------------

New functions:

- Add update.list() function to replace named elements a list.

- Add mv() function to rename an object.

- Add first(), last(), first<-() and last<-() functions to extract
  or replace first or last vector/list elements.

New features:

- Add 'byrow' argument to lowerTriangle() and upperTriangle() functions.

Other changes:

- humanReadable() now properly handles a single value for the argument
  'justify'.

- Improve logging and error reporting for remove.vars().

- read.xls() now handles latin-1 files properly on MS Windows.

- write.fwf() now properly handles matrix arguments.




Changes in 2.17.0 (2015-07-02)
------------------------------

New features:

- Add new argument 'byrow' to upperTriangle(), lowerTriangle(),
  upperTriangle<-(), and lowerTriangle<-() to specify by row rather than
  by column order.  This makes it simpler to copy values between the
  lower and upper triangular areas, e.g. to construct a symmetric
  matrix.

Other changes:

- Add inline comments to tests to alert reviewers of expected diffs on
  systems lacking the libraries for read.xls() to support XLSX
  formatted files.




Changes in 2.16.1 (2015-04-28)
-----------------------------

Bug fixes:

- mapLevels() no longer generates warnings about conversion of lists to vectors.

Other changes:

- Requirement for Perl version 5.10.0 or later is specified in the
  package DESCRITION.

- first() and last() are now simply wrappers for calls to 'head(x, n=1)' and
  'tail(x, n=1)', respectively.




Changes in 2.16.0 (2015-04-25)
------------------------------

New features:

- New functions first() and last() to return the first or last element
  of an object.

- New functions left() and right() to return the leftmost or
  rightmost n (default to 6) columns of a matrix or data frame.

- New 'scientific' argument to write.fwf().  Set 'scientific=FALSE' to
  prevent numeric columns from being displayed using scientific
  notification.

- The 'standard' argument to humanReadable() now accepts three values,
  'SI' for base 1000 ('MB'), 'IEC' for base 1024 ('MiB'), and 'Unix'
  for base 1024 and single-character units ('M')

- object.size() now returns objects with S3 class 'object_sizes' (note the
  final 's') to avoid conflicts with methods in utils for class 'object_size'
  provided by package 'utils' which can only handle a scalar size.

- New 'units' argument to humanReadable()--and hence to
  print.object_sizes() and format.object_sizes()--that permits
  specifying the unit to use for all values.  Use 'bytes' to display
  all values with the unit 'bytes', use 'auto' (or leave it missing) to
  automatically select the best unit, and use a unit from the selected
  standard to use that unit (i.e. 'MiB').

- The default arguments to humanReadable() have changed. The defaults
  are now 'width=NULL' and 'digits=1', so that the default behavior is
  now to show digit after the decimal for all values.

Bug fixes:

- reorder.factor() was ignoring the argument 'X' unless 'FUN' was
  supplied, making it incompatible with the behavior of
  stats:::reorder.default(). This has been corrected, so that calling
  reorder on a factor with arguments 'X' and/or 'FUN' should now
  return the same results whether gdata is loaded or not.  (Reported
  by Sam Hunter.)

- write.fwf() now properly supports matrix objects, including matrix
  objects without column names. (Reported by Carl Witthoft.)

Other changes:

- Replaced deprecated PERL function POSIX::isdigit in xls2csv.pl
  (which is used by read.xls() ) with an equivalent regular
  expression. (Reported by both Charles Plessy, Gerrit-jan Schutten,
  and Paul Johnson.  Charles also provided a patch to correct the
  issue.)

- aggregate.table(), which has been defunct gdata 2.13.3 (2014-04-04)
  has now been completely removed.




Changes in 2.14.0 (2014-08-27)
------------------------------

Bug Fixes:

- read.xls() can now properly process XLSX files with up to 16385 columns (the
  maximum generated by Microsoft Excel).

- read.xls() now properly handles XLS/XLSX files that use 1904-01-01
  as the reference value for dates instead of 1900-01-01 (the
  default for MS-Excel files created on the Mac).

Other changes:

- Updated perl libraries and code underlying read.xls() to the latest
  version, including switching from Spreadsheet::XLSX to
  Spreadsheet::ParseXLSX.




Changes in 2.13.3 (2014-04-04)
------------------------------

Bug Fixes:

- Unit tests were incorrectly checking for equality of optional POSIXlt
  components.  (Bug reported by Brian Ripley).

Other Changes:

- 'aggregate.table' is now defunct.  See '?gdata-defunct' for details.

- Unit tests and vignettes now follow R standard practice.

- Minor changes to clean up R CMD check warnings.




Changes in 2.13.2 (2013-06-28)
------------------------------

Enhancements:

- Simplify ll() by converting a passed list to an environment,
  avoiding the need for special casing and the use of attach/detach.

- Working of deprecation warning message in aggregate.table clarified.




Changes in 2.13.1 (2013-03-24)
------------------------------

Enhancements:

- Replaced calls to depreciated function ".path.package" with the new
  public function "path.package".




Changes in 2.13.0 (2012-09-20)
-----------------------------

New features:

- New 'duplicated2' function which returns TRUE for *all* elements
  that are duplicated, including the first, contributed by Liviu
  Andronic.  This differs from 'duplicated', which only returns the
  second and following (second-to last and previous when
  'fromLast=TRUE') duplicate elements.

- New 'ans' functon to return the value of the last evaluated
  top-level function (a convenience function for accessing
  .Last.value), contributed by Liviu Andonic.

Bug Fixes:

- On windows, warning messages printed to stdout by perl were being
  included in the return value from 'system', resulting in errors in
  'sheetCount' and 'sheetNames'.  Corrected.

- The 'MedUnits' column names 'SIUnits' and 'ConventionalUnits' were
  reversed and misspelled.




Changes in 2.12.0 (2012-09-12)
------------------------------

Other Changes:

- 'stats::aggregate' was made into a generic on 27-Jan-2010, so that
   attempting to call 'aggregate' on a 'table' object will now
   incorrectly call 'aggregate.table'.  Since 'aggregate.table' can be
   replaced by a call to tapply using two index vectors, e.g.
       aggregate.table(x, by1=a, by2=b, mean)
   can be replaced by
       tapply(x, INDEX=list(a, b), FUN=mean),
   the 'aggregate.table' function will now display a warning that it
   is depreciated and recommending the equivalent call to tapply. It
   will be removed entirely in a future version of gdata.




Changes in 2.11.1 (2012-08-22)
------------------------------

Enhancements:

- read.xls() now supports fileEncoding argument to allow non-ascii
  encoded data to be handled.  See the manual page for an example.

Bug Fixes:

- The perl script utilized by read.xls() was incorrectly appending a
  space character at the end of each line, causing problems with
  character and NA entries in the final column.




Changes in 2.11.0 (2012-06-18)
------------------------------

New Features:

- read.xls() and supporting functions now allow blank lines to be
  preserved, rather than skipped, by supplying the argument
  "blank.lines.skip=FALSE". The underlying perl function has been
  extended to suppor this via an optional "-s" argument which, when
  present, *preserves* blank lines during the conversion.  (The
  default behavior remains unchanged.)

Other Changes:

- Add SystemRequirements field specifying that perl is necessary for
  gdata to function fully.




Changes in 2.10.6 (2012-06-12)
------------------------------

Bug fixes:

- gdata::nobs.default() needs to handle logical vectors in addition to
  numeric vectors.




Changes in 2.10.5 (2012-06-08)
------------------------------

Bug fixes:

- Mark example for installXLSsupport() as dontrun so R CMD check won't
  fail on systems where PERL is not fully functional.

- Correct name of installXLSsupport() in tests/test.read.xls.R.

Other Changes:

- Add dependency on R 2.13.0, since that is when stats::nobs appeared.




Changes in 2.10.2 (2012-06-06)
------------------------------

Bug fixes:

- Fix issues in nobs.default identified in testing with the gmodels package.




Changes in 2.10.1 (2012-06-06)
------------------------------

Bug fixes:

- Undo removal of 'nobs' and 'nobs.lm'.  Instead define aliases for
  'nobs' and 'nobs.lm' to support backward compatibility for packages
  depending on gdata.




Changes in 2.10.0 (2012-06-05)
------------------------------

New features:

- New ls.funs() function to list all objects of class function in the
  specified environment.

- New startsWith() function to determine if a string "starts with" the
  specified characters.

Enhancements:

- Add 'na.strings' argument to read.xls() to convert Excel's '#DIV/0!' to NA.

Bug fixes:

- Correct various R CMD check warnings

Other changes:

- Base S3 method for nobs() and nobs.lm() method removed since these
  are now provided in the stats package.




Changes in 2.9.0 (2011-09-30)
-----------------------------

New features:

- Add centerText() function to center text strings for a specified
  width.

- Add case() function, a vectorized variant of the base::switch()
  function, which is useful for converting numeric codes into factors.

Enhancements:

- Minor improvements to xls2csv() man page.




Changes in 2.8.1 (2011-04-15)
-----------------------------

Enhancements:

- nPairs() gains a summary method that shows how many times each variable
  is known, while the other variable of a pair is not

Bug fixes:

- Fix errors on windows when R or Perl install path includes spaces by properly
  quoting the path.




Changes in 2.8.1 (2010-11-12)
-----------------------------

Enhancements:

- Minor improvement to Args(), read.xls() man page.

Bug fixes:

- Modify write.fwf() to capture and pass on additional arguments for
  write.table(). This resolves a bug reported by Jan Wijffels.

- Modify xls2sep.R to avoid use of file.access() which is unreliable
  on Windows network shares.




Changes in 2.8.0 (2010-04-03)
-----------------------------

Enhancements:

- When loaded, gdata (via an .onAttach() function) now checks:
   1) if perl is available
   2) whether the perl libraries for XLS support are available
   3) whether the perl libraries for XLSX support are available

  If perl is not available, an appropriate warning message is displayed.

  If necessary perl libraries are not available, a warning message is
  displayed, as is a message suggesting the user run the (new)
  installXLSXsupport() function to attempt to install the necessary
  perl libraries.

- The function installXLSXsupport() has been provided to install the
  binary perl modules that read.xls needs to support Excel 2007+
  'XLSX' files.




Changes in 2.7.3 (2010-04-02)
-----------------------------

Enhancements:

- New xlsFormats() command to determine which Excel formats are
  supported (XLS, XLSX).

Bug Fixes:

- No longer attempt to install perl modules Compress::Raw::Zlib and
  Spreadsheet::XLSX at build/compile time.  This should resolve recent
  build issues, particularly on Windows.

- All perl code can now operate (but generate warnings) when perl modules
  Compress::Raw::Zlib and Spreadsheet::XLSX when are not installed.

- Also update Greg's email address.




Changes in 2.7.1 (2010-02-19)
-----------------------------

Enhancements:

- on Windows attempts to locate ActiveState perl if perl= not specified and
  Rtools perl would have otherwise been used in read.xls and other perl
  dependent functions.




Changes in 2.7.0 (2010-01-25)
-----------------------------

Bug Fixes:

- Fix building of Perl libraries on Win32




Changes in 2.7.0 (2010-01-25)
-----------------------------

Enhancements:

- read.xls() now supports Excel 2007 'xlsx' files.

- read.xls() now allows specification of worksheet by name

- read.xls() now supports ftp URLs.

- Improved ll() so user can limit output to specified classes

New Functions:

- sheetCount() and sheetNames() to determine the number and names of
  worksheets in an Excel file, respectively.

Bug Fixes:

- Fix formatting warning in frameApply().

- Resolve crash of "ll(.GlobalEnv)"




Changes in 2.6.1 (2009-07-15)
-----------------------------

Bug Fixes

- Modify unit tests to avoid issues related to time zones.




Changes in 2.6.0 (2009-07-15)
-----------------------------

Bug Fixes

- Correct minor typos & issues in man pages for write.fwf(),
  resample() (Greg Warnes)

- Correct calculation of object sizes in env() and ll() (Gregor Gorjanc)

New Features

- Add support for using tab for field separator during translation
  from xls format in read.xls (Greg Warnes)

- Enhanced function object.size that returns the size of multiple
  objects. There is also a handy print method that can print size of
  an object in "human readable" format when
    options(humanReadable=TRUE)
  or
    print(object.size(x), humanReadable=TRUE).
  (Gregor Gorjanc)

- New function wideByFactor that reshapes given dataset by a given
  factor - it creates a "multivariate" data.frame. (Gregor Gorjanc)

- New function nPairs that gives the number of variable pairs in a
  data.frame or a matrix. (Gregor Gorjanc)

- New functions getYear, getMonth, getDay, getHour, getMin, and getSec
  for extracting the date/time parts from objects of a date/time
  class. (Gregor Gorjanc)

- New function bindData that binds two data frames into a multivariate
  data frame in a different way than merge. (Gregor Gorjanc)

Other Changes

- Correct Greg's email address




Changes in 2.5.0 (2008-12-31)
-----------------------------

- New function .runRUnitTestsGdata that enables run of all RUnit tests during
  the R CMD check as well as directly from within R.

- Enhanced function object.size that returns the size of multiple objects.
  There is also a handy print method that can print size of an object in
  "human readable" format when options(humanReadable=TRUE) or
  print(x, humanReadable=TRUE).

- New function bindData that binds two data frames into a multivariate data
  frame in a different way than merge.

- New function wideByFactor that reshapes given dataset by a given factor -
  it creates a "multivariate" data.frame.

- New functions getYear, getMonth, getDay, getHour, getMin, and getSec for
  extracting the date/time parts from objects of a date/time class.

- New function nPairs that gives the number of variable pairs in a data.frame
  or a matrix.

- New function trimSum that sums trimmed values.

- New function cbindX that can bind objects with different number of rows.

- write.fwf gains the width argument. The value for unknown can increase or
  decrease the width of the columns. Additional tests and documentation fixes.




Changes in 2.4.2 (2008-05-11)
-----------------------------

- Enhancements and bug fixes for read.xls() and xls2csv():

    - More informative log messages when verbose=TRUE

    - File paths containing spaces or other non-traditional characters
      are now properly handled

    - Better error messages, particularly when perl fails to generate
      an output .csv file.

    - The 'shortcut' character "~" (meaning user's home directory) is
      now properly handled in file paths.

    - XLS files created by OpenOffice are now properly handled.  Thanks to
       Robert Burns for pointing out the patch
       (http://rt.cpan.org/Public/Bug/Display.html?id=7206)




Changes in 2.4.1 (2008-03-24)
-----------------------------

- Update perl libraries needed by xls2csv() and read.xls()
  to latest available versions on CRAN.

- Add read.xls() to exported function list

- Correct iris.xls example file.  It didn't contain the complete
  & properly formatted iris data set. Fixed.

- Fix typo in win32 example for read.xls()




Changes in 2.4.0 (2008-01-30)
-----------------------------

- The keep() function now includes an 'all' argument to specify how
  objects with names starting with '.' are handled.

- keep() now shows an informative warning message when a requested
  object does not exist

- New vignette "Mapping Levels of a Factor" describing the use of
  mapLevels().

- New vignette "Working with Unknown Values" describing the use of
  isUnknown() and unknownToNA().

- Several enhancements to read.xls() (thanks to Gabor Grothendieck):

  - New function xls2csv(), which handles converting an xls file
    to a csv file and returns a connection to the
    temporary csv file

  - xls2csv() and read.xls() both allow a file or a url to be specified

  - read.xls() has a new 'pattern' argument which, if supplied,
    will ignore everything prior to the first line in th csv file
    that matches the pattern.  This is typically used if there
    are a variable number of comment lines prior to the header
    in which case one can specify one of the column
    headings as the pattern.  read.xls should
    be compatible with the old read.xls.

- Minor fixes to drop.levels(), is.what().

- Implementation of unit tests for most functions.




Changes in 2.3.1 (2006-10-29)
-----------------------------

- Arguments as well as their position of reorder.factor have been changed
  to conform with reorder.factor method in stats package, due to collision
  bug. Argument 'make.ordered' is now 'order' and old argument 'order' is
  now 'new.order'! Therefore, you have to implicitly specify new.order i.e.

  reorder(trt, new.order=c("PLACEBO", "300 MG", "600 MG", "1200 MG"))

- trim() gains ... argument.

- Added "unknown" methods for matrices.

- Added c() method for factors based on mapLevels() functions.

- Added write.fwf, which writes file in *F*ixed *W*idth *F*ormat.




Changes in 2.3.0 (2006-09-19)
-----------------------------

- Added mapLevels(), which produces a map with information on levels and/or
  internal integer codes.  Contributed by Gregor Gorjanc.

- Extended dropLevels() to work on the factors contained in a data
  frame, as well as individual factors.

- Add unknown(), which changes given unknown value to NA and vice
  versa. Contributed by Gregor Gorjanc.

- Extended trim() to handle a variety of data types data.frames,
  lists, factors, etc.  Code changes contributed by Gregor Gorjanc.

- Added resample() command that acts like sample() except that it
  _always_ samples from the arguments provided, even if only a single
  argument is present.  This differs from sample() which behaves
  differently in this case.

- Updated my email address.




Changes in 2.1.2 (2012-10-29)
-----------------------------

 - Fixed bug in interleave.R - option to covert 1-column matrices to
   vector (based on Andrew Burgess's suggestion)

 - Updated Greg and Jim's email adresses

 - ll.R: Suppressed warning message in attach() call.

 - frameApply.Rd, reorder.Rd: Remove explicit loading of
   gtools in examples, so that failure to import functions from
   gtools gets properly caught by running the examples.

 - upperTriangle.R, man/upperTriangle.Rd: Add functions for
   extracting and modifying the upper and lower trianglular components of
   matrices.

 - is.what.R: Replaced the "not.using" vector with a more robust
   try(get(test)) to find out whether a particular is.* function
   returns a logical of length one.

-  DESCRIPTION: Added Suggests field

-  Updated the example in frameApply




Changes in 2.0.8 (2005-06-02)
-----------------------------

 - Added DESCRIPTION and removed DESCRIPTION.in

 - Updated ll.Rd documentation

 - Fixed bug in Args.R, is.what.R, ll.R