mark 0.5.2

New features

• normalize() added to normalize values in vectors, matrices, and data.frames by specified ranges and boundaries #143
• get_labels() and other label related functions now get exact matches for "label" attributes #141
• recode_only(), and recode_by() now accept a named list() for by #96]
• switch_in_case() now handles functions in rh
• update_version() now correctly checks result of embedded utils::menu() call for updating the version #123
• require_namespace() now accepts multiple namespaces #121
• unique.fact() S3 method #86
• recode_only() and recode_by() can accept a single value for val #72
• fact_reverse() for reversing fact levels #78
• as.Date.fact() added #108
• as.character.fact() added
• [.fact added
• read_bib() better handles fields where = is present in the text #117

Fixes

• fact.haven_labelled() works properly and retains the label attribute #136
• drop_levels() is exported #105
• recode_by() and recode_only() handle factors better #81
• Functions that made use of shell.exec() now try to determine the appropriate method of opening a file base on OS. #126
• Internal functions for potentially coercing factor levels into dates no longer try to check for "%Z" in the date format #147

Breaking changes

• reverse() has been removed (use flip() instead)
• assign_label() has been removed (use assign_labels() instead)
• percentile_rank(times) is deprecated in favor of percent_rank(weights)

Non visible changes

• print.fact() rewritten as a slightly modified print.factor() #109
• percentile_rank() improvements #131