labelled 2.13.0

New features

• add a null_action argument to val_labels(), val_label() and a
  .null_action argument to set_value_labels(), add_value_labels() and
  remove_value_labels() (#145)
• new functions update_variable_labels_with() and update_value_labels_with()
  allowing to update variable/value labels with a custom function (#153)

Bug fix

• avoid an error with print.look_for() when console pane is physically shrunk
  too small (#148)
• fix in recode.haven_labelled() when .x contains NA and
  .combine_value_labels = TRUE (#151)
• produce an error when trying to assign value labels to a Date vector (#156)

labelled 2.12.0

New features

• support of variable labels for packed columns, see dedicated vignette (#142)
• new helpers label_attribute(), get_label_attribute() and
  set_label_attribute() to manipulate the "label" attribute on any
  object (#142)
• new functions get_variable_labels(), get_value_labels(), get_na_values()
  and get_na_range() identical to var_label(), val_labels(), na_values()
  and na_range(), respectively
• to_character() method for data frames (#140)

labelled 2.11.0

Improvements

• set_value_labels(), add_value_labels(), remove_value_labels(),
  set_variable_labels(), set_na_range() and set_na_values() can now be
  applied on a vector (#126)
• new argument null_action for var_label() when applied on a
  data frame (#131)
• look_for() now returns "missing" (number of NAs) by default (#133)

Bug fixes

• bug fix in print.look_for() (#135)
• bug fix in unlabelled() for classic vectors, now remained unchanged (#137)

labelled 2.10.0

• look_for() now accepts survey objects (#121)

labelled 2.9.1

• improved error messages for missing variable names (#118, @ajb5d)
• better implementation of look_for() when no keyword is provided (#116)
• bug fix in user_na_to_tagged_na() (#114)

labelled 2.9.0

look_for() improvements:

• new function look_for_and_select() (#87)
• look_for() can now search within factor levels and value labels (#104)

improvements for tagged NAs:

• better printing of value labels (#89)
• new functions user_na_to_tagged_na(), tagged_na_to_user_na() and
  tagged_na_to_regular_na()
• new option explicit_tagged_na in to_factor() and to_character()
• new functions unique_tagged_na(), duplicated_tagged_na(),
  order_tagged_na(), sort_tagged_na() (#90, #91)

other improvements:

• new functions is_user_na() and is_regular_na()
• new set of unit tests (#99)
• trying to apply a value label, na_range() or na_values() to a factor
  will now produce an error
• bug fix in foreign_to_labelled() for Stata files (#100)

labelled 2.8.0

• new helper recode_if() for recoding values based on condition, variable
  and value labels being preserved (#82)
• look_for() could be time consuming for big data frames. Now, by default,
  only basic details of each variable are computed. You can compute all details
  with details = "full" (#77)
• printing of look_for() results has been updated and do not rely anymore
  on pillar (#85)
• to_labelled() can properly manage factors whose levels are coded as
  "[code] level", as produced by to_factor(levels = "prefixed")
  (#74 @courtiol)
• new function is_prefixed() to check if a factor is prefixed
• bug fix for na_range<- and na_values<- when applied to a data.frame (#80)

labelled 2.7.0

• a .values argument has been added to set_na_values() and set_na_range(), allowing to pass a list of values
• a .strict option has been added to set_variable_labels(),
  set_value_labels(), add_value_labels(), remove_value_labels(),
  set_na_values() and set_na_range(),
  allowing to pass values for columns not observed in the data (it could be useful
  for using a same list of labels for several data.frame sharing some
  variables) (#70)
• copy_labels() is less restrictive for non labelled vectors, copying variable label
  even if the two vectors are not of the same type (#71)
• a .strict option has been added to copy_labels() (#71)

labelled 2.6.0

• look_for() has been redesigned:
  • look_for() now returns a tibble
  • columns with multiple values for each variable are now stored as named lists
  • a print method has been added for a clearer presentation of results in the console
  • use lookfor_to_long_format() to convert results with one row per factor level
    and per value label
  • use convert_list_columns_to_character() to convert list columns to
    simpler character vectors
  • generate_dictionary() is an equivalent of look_for()
• set_variable_labels, set_value_labels, add_value_labels,
  and remove_value_labels now accept "tidy dots" (#67 @psanker)
• new function names_prefixed_by_values() to get the names of a vector
  prefixed by their corresponding value

labelled 2.5.0

• new .keep_value_labels argument for recode.haven_labelled()
• new .combine_value_labels argument for recode.haven_labelled() (#61)
• new drop_unused_value_labels() method
• an additional .labels argument for set_value_labels()
• user_na_to_na argument has been added to to_character.haven_labelled()
• %>% is now imported from dplyr
• a cheatsheet has been added (#47)
• internal documentation is now using roxygen2 markdown support