ggplot2 3.3.3

This is a small patch release mainly intended to address changes in R and CRAN.
It further changes the licensing model of ggplot2 to an MIT license.

• Update the ggplot2 licence to an MIT license (#4231, #4232, #4233, and #4281)
• Use vdiffr conditionally so ggplot2 can be tested on systems without vdiffr
• Update tests to work with the new all.equal() defaults in R >4.0.3

ggplot2 3.3.2

This is a small release focusing on fixing regressions introduced in 3.3.1.

• Added an outside option to annotation_logticks() that places tick marks
  outside of the plot bounds. (#3783, @kbodwin)

• annotation_raster() adds support for native rasters. For large rasters,
  native rasters render significantly faster than arrays (@kent37, #3388)

• Facet strips now have dedicated position-dependent theme elements
  (strip.text.x.top, strip.text.x.bottom, strip.text.y.left,
  strip.text.y.right) that inherit from strip.text.x and strip.text.y,
  respectively. As a consequence, some theme stylings now need to be applied to
  the position-dependent elements rather than to the parent elements. This
  change was already introduced in ggplot2 3.3.0 but not listed in the
  changelog. (@thomasp85, #3683)

• Facets now handle layers containing no data (@yutannihilation, #3853).

• A newly added geom geom_density_2d_filled() and associated stat
  stat_density_2d_filled() can draw filled density contours
  (@clauswilke, #3846).

• A newly added geom_function() is now recommended to use in conjunction
  with/instead of stat_function(). In addition, stat_function() now
  works with transformed y axes, e.g. scale_y_log10(), and in plots
  containing no other data or layers (@clauswilke, #3611, #3905, #3983).

• Fixed a bug in geom_sf() that caused problems with legend-type
  autodetection (@clauswilke, #3963).

• Support graphics devices that use the file argument instead of fileneame
  in ggsave() (@bwiernik, #3810)

• Default discrete color scales are now configurable through the options() of
  ggplot2.discrete.colour and ggplot2.discrete.fill. When set to a character
  vector of colour codes (or list of character vectors) with sufficient length,
  these colours are used for the default scale. See help(scale_colour_discrete)
  for more details and examples (@cpsievert, #3833).

• Default continuous colour scales (i.e., the options()
ggplot2.continuous.colour and ggplot2.continuous.fill, which inform the
  type argument of scale_fill_continuous() and scale_colour_continuous())
  now accept a function, which allows more control over these default
  continuous_scale()s (@cpsievert, #3827).

• A bug was fixed in stat_contour() when calculating breaks based on
  the bins argument (@clauswilke, #3879, #4004).

• Data columns can now contain Vector S4 objects, which are widely used in the
  Bioconductor project. (@teunbrand, #3837)

ggplot2 3.3.1

This is a small release with no code change. It removes all malicious links to a
site that got hijacked from the readme and pkgdown site.

ggplot2 3.3.0

This is a minor release but does contain a range of substantial new features,
along with the standard bug fixes. The release contains a few visual breaking
changes, along with breaking changes for extension developers due to a shift in
internal representation of the position scales and their axes. No user breaking
changes are included.

This release also adds Dewey Dunnington (@paleolimbot) to the core team.

Breaking changes

There are no user-facing breaking changes, but a change in some internal
representations that extension developers may have relied on, along with a few
breaking visual changes which may cause visual tests in downstream packages to
fail.

• The panel_params field in the Layout now contains a list of list of
  ViewScale objects, describing the trained coordinate system scales, instead
  of the list object used before. Any extensions that use this field will likely
  break, as will unit tests that checks aspects of this.

• element_text() now issues a warning when vectorized arguments are provided,
  as in colour = c("red", "green", "blue"). Such use is discouraged and not
  officially supported (@clauswilke, #3492).

• Changed theme_grey() setting for legend key so that it creates no border
  (NA) rather than drawing a white one. (@annennenne, #3180)

• geom_ribbon() now draws separate lines for the upper and lower intervals if
  colour is mapped. Similarly, geom_area() and geom_density() now draw
  the upper lines only in the same case by default. If you want old-style full
  stroking, use outline.type = "full" (@yutannihilation, #3503 / @thomasp85, #3708).

New features

• The evaluation time of aesthetics can now be controlled to a finer degree.
  after_stat() supersedes the use of stat() and ..var..-notation, and is
  joined by after_scale() to allow for mapping to scaled aesthetic values.
  Remapping of the same aesthetic is now supported with stage(), so you can
  map a data variable to a stat aesthetic, and remap the same aesthetic to
  something else after statistical transformation (@thomasp85, #3534)

• All coord_*() functions with xlim and ylim arguments now accept
  vectors with NA as a placeholder for the minimum or maximum value
  (e.g., ylim = c(0, NA) would zoom the y-axis from 0 to the
  maximum value observed in the data). This mimics the behaviour
  of the limits argument in continuous scale functions
  (@paleolimbot, #2907).

• Allowed reversing of discrete scales by re-writing get_limits()
  (@AnneLyng, #3115)

• All geoms and stats that had a direction (i.e. where the x and y axes had
  different interpretation), can now freely choose their direction, instead of
  relying on coord_flip(). The direction is deduced from the aesthetic
  mapping, but can also be specified directly with the new orientation
  argument (@thomasp85, #3506).

• Position guides can now be customized using the new guide_axis(), which can
  be passed to position scale_*() functions or via guides(). The new axis
  guide (guide_axis()) comes with arguments check.overlap (automatic removal
  of overlapping labels), angle (easy rotation of axis labels), and
  n.dodge (dodge labels into multiple rows/columns) (@paleolimbot, #3322).

• A new scale type has been added, that allows binning of aesthetics at the
  scale level. It has versions for both position and non-position aesthetics and
  comes with two new guides (guide_bins and guide_coloursteps)
  (@thomasp85, #3096)

• scale_x_continuous() and scale_y_continuous() gains an n.breaks argument
  guiding the number of automatic generated breaks (@thomasp85, #3102)

• Added stat_contour_filled() and geom_contour_filled(), which compute
  and draw filled contours of gridded data (@paleolimbot, #3044).
  geom_contour() and stat_contour() now use the isoband package
  to compute contour lines. The complete parameter (which was undocumented
  and has been unused for at least four years) was removed (@paleolimbot, #3044).

• Themes have gained two new parameters, plot.title.position and
  plot.caption.position, that can be used to customize how plot
  title/subtitle and plot caption are positioned relative to the overall plot
  (@clauswilke, #3252).

Extensions

• Geom now gains a setup_params() method in line with the other ggproto
  classes (@thomasp85, #3509)

• The newly added function register_theme_elements() now allows developers
  of extension packages to define their own new theme elements and place them
  into the ggplot2 element tree (@clauswilke, #2540).

Minor improvements and bug fixes

• coord_trans() now draws second axes and accepts xlim, ylim,
  and expand arguments to bring it up to feature parity with
  coord_cartesian(). The xtrans and ytrans arguments that were
  deprecated in version 1.0.1 in favour of x and y
  were removed (@paleolimbot, #2990).

• coord_trans() now calculates breaks using the expanded range
  (previously these were calculated using the unexpanded range,
  which resulted in differences between plots made with coord_trans()
  and those made with coord_cartesian()). The expansion for discrete axes
  in coord_trans() was also updated such that it behaves identically
  to that in coord_cartesian() (@paleolimbot, #3338).

• expand_scale() was deprecated in favour of expansion() for setting
  the expand argument of x and y scales (@paleolimbot).

• geom_abline(), geom_hline(), and geom_vline() now issue
  more informative warnings when supplied with set aesthetics
  (i.e., slope, intercept, yintercept, and/or xintercept)
  and mapped aesthetics (i.e., data and/or mapping).

• Fix a bug in geom_raster() that squeezed the image when it went outside
  scale limits (#3539, @thomasp85)

• geom_sf() now determines the legend type automatically (@microly, #3646).

• geom_sf() now removes rows that can't be plotted due to NA aesthetics
  (#3546, @thomasp85)

• geom_sf() now applies alpha to linestring geometries
  (#3589, @yutannihilation).

• gg_dep() was deprecated (@perezp44, #3382).

• Added function ggplot_add.by() for lists created with by(), allowing such
  lists to be added to ggplot objects (#2734, @Maschette)

• ggplot2 no longer depends on reshape2, which means that it no longer
  (recursively) needs plyr, stringr, or stringi packages.

• Increase the default nbin of guide_colourbar() to place the ticks more
  precisely (#3508, @yutannihilation).

• manual_scale() now matches values with the order of breaks whenever
  values is an unnamed vector. Previously, unnamed values would match with
  the limits of the scale and ignore the order of any breaks provided. Note
  that this may change the appearance of plots that previously relied on the
  unordered behaviour (#2429, @idno0001).

• scale_manual_*(limits = ...) now actually limits the scale (#3262,
  @yutannihilation).

• Fix a bug when show.legend is a named logical vector
  (#3461, @yutannihilation).

• Added weight aesthetic option to stat_density() and made scaling of
  weights the default (@annennenne, #2902)

• stat_density2d() can now take an adjust parameter to scale the default
  bandwidth. (#2860, @haleyjeppson)

• stat_smooth() uses REML by default, if method = "gam" and
  gam's method is not specified (@ikosmidis, #2630).

• stacking text when calculating the labels and the y axis with
  stat_summary() now works (@ikosmidis, #2709)

• stat_summary() and related functions now support rlang-style lambda functions
  (#3568, @dkahle).

• The data mask pronoun, .data, is now stripped from default labels.

• Addition of partial themes to plots has been made more predictable;
  stepwise addition of individual partial themes is now equivalent to
  addition of multple theme elements at once (@clauswilke, #3039).

• Facets now don't fail even when some variable in the spec are not available
  in all layers (@yutannihilation, #2963).

ggplot2 3.2.1

This is a patch release fixing a few regressions introduced in 3.2.0 as well as
fixing some unit tests that broke due to upstream changes.

• position_stack() no longer changes the order of the input data. Changes to
  the internal behaviour of geom_ribbon() made this reordering problematic
  with ribbons that spanned y = 0 (#3471)
• Using qplot() with a single positional aesthetic will no longer title the
  non-specified scale as "NULL" (#3473)
• Fixes unit tests for sf graticule labels caused by chages to sf

ggplot2 3.2.0

This is a minor release with an emphasis on internal changes to make ggplot2
faster and more consistent. The few interface changes will only affect the
aesthetics of the plot in minor ways, and will only potentially break code of
extension developers if they have relied on internals that have been changed.
This release also sees the addition of Hiroaki Yutani (@yutannihilation) to the
core developer team.

With the release of R 3.6, ggplot2 now requires the R version to be at least 3.2,
as the tidyverse is committed to support 5 major versions of R.

Breaking changes

• Two patches (#2996 and #3050) fixed minor rendering problems. In most cases,
  the visual changes are so subtle that they are difficult to see with the naked
  eye. However, these changes are detected by the vdiffr package, and therefore
  any package developers who use vdiffr to test for visual correctness of ggplot2
  plots will have to regenerate all reference images.

• In some cases, ggplot2 now produces a warning or an error for code that previously
  produced plot output. In all these cases, the previous plot output was accidental,
  and the plotting code uses the ggplot2 API in a way that would lead to undefined
  behavior. Examples include a missing group aesthetic in geom_boxplot() (#3316),
  annotations across multiple facets (#3305), and not using aesthetic mappings when
  drawing ribbons with geom_ribbon() (#3318).

New features

• This release includes a range of internal changes that speeds up plot
  generation. None of the changes are user facing and will not break any code,
  but in general ggplot2 should feel much faster. The changes includes, but are
  not limited to:

  • Caching ascent and descent dimensions of text to avoid recalculating it for
    every title.

  • Using a faster data.frame constructor as well as faster indexing into
    data.frames

  • Removing the plyr dependency, replacing plyr functions with faster
    equivalents.

• geom_polygon() can now draw polygons with holes using the new subgroup
  aesthetic. This functionality requires R 3.6.0 (@thomasp85, #3128)

• Aesthetic mappings now accept functions that return NULL (@yutannihilation,
  #2997).

• stat_function() now accepts rlang/purrr style anonymous functions for the
  fun parameter (@dkahle, #3159).

• geom_rug() gains an "outside" option to allow for moving the rug tassels to
  outside the plot area (@njtierney, #3085) and a length option to allow for
  changing the length of the rug lines (@daniel-wells, #3109).

• All geoms now take a key_glyph paramter that allows users to customize
  how legend keys are drawn (@clauswilke, #3145). In addition, a new key glyph
  timeseries is provided to draw nice legends for time series
  (@mitchelloharawild, #3145).

Extensions

• Layers now have a new member function setup_layer() which is called at the
  very beginning of the plot building process and which has access to the
  original input data and the plot object being built. This function allows the
  creation of custom layers that autogenerate aesthetic mappings based on the
  input data or that filter the input data in some form. For the time being, this
  feature is not exported, but it has enabled the development of a new layer type,
  layer_sf() (see next item). Other special-purpose layer types may be added
  in the future (@clauswilke, #2872).

• A new layer type layer_sf() can auto-detect and auto-map sf geometry
  columns in the data. It should be used by extension developers who are writing
  new sf-based geoms or stats (@clauswilke, #3232).

• x0 and y0 are now recognized positional aesthetics so they will get scaled
  if used in extension geoms and stats (@thomasp85, #3168)

• Continuous scale limits now accept functions which accept the default
  limits and return adjusted limits. This makes it possible to write
  a function that e.g. ensures the limits are always a multiple of 100,
  regardless of the data (@econandrew, #2307).

Minor improvements and bug fixes

• cut_width() now accepts ... to pass further arguments to base::cut.default()
  like cut_number() and cut_interval() already did (@cderv, #3055)

• coord_map() now can have axes on the top and right (@karawoo, #3042).

• coord_polar() now correctly rescales the secondary axis (@linzi-sg, #3278)

• coord_sf(), coord_map(), and coord_polar() now squash -Inf and Inf
  into the min and max of the plot (@yutannihilation, #2972).

• coord_sf() graticule lines are now drawn in the same thickness as panel grid
  lines in coord_cartesian(), and seting panel grid lines to element_blank()
  now also works in coord_sf()
  (@clauswilke, #2991, #2525).

• economics data has been regenerated. This leads to some changes in the
  values of all columns (especially in psavert), but more importantly, strips
  the grouping attributes from economics_long.

• element_line() now fills closed arrows (@yutannihilation, #2924).

• Facet strips on the left side of plots now have clipping turned on, preventing
  text from running out of the strip and borders from looking thicker than for
  other strips (@karawoo, #2772 and #3061).

• ggplot2 now works in Turkish locale (@yutannihilation, #3011).

• Clearer error messages for inappropriate aesthetics (@clairemcwhite, #3060).

• ggplot2 no longer attaches any external packages when using functions that
  depend on packages that are suggested but not imported by ggplot2. The
  affected functions include geom_hex(), stat_binhex(),
  stat_summary_hex(), geom_quantile(), stat_quantile(), and map_data()
  (@clauswilke, #3126).

• geom_area() and geom_ribbon() now sort the data along the x-axis in the
  setup_data() method rather than as part of draw_group() (@thomasp85,
  #3023)

• geom_hline(), geom_vline(), and geom_abline() now throw a warning if the
  user supplies both an xintercept, yintercept, or slope value and a
  mapping (@RichardJActon, #2950).

• geom_rug() now works with coord_flip() (@has2k1, #2987).

• geom_violin() no longer throws an error when quantile lines fall outside
  the violin polygon (@thomasp85, #3254).

• guide_legend() and guide_colorbar() now use appropriate spacing between legend
  key glyphs and legend text even if the legend title is missing (@clauswilke, #2943).

• Default labels are now generated more consistently; e.g., symbols no longer
  get backticks, and long expressions are abbreviated with ...
  (@yutannihilation, #2981).

• All-Inf layers are now ignored for picking the scale (@yutannihilation,
  #3184).

• Diverging Brewer colour palette now use the correct mid-point colour
  (@dariyasydykova, #3072).

• scale_color_continuous() now points to scale_colour_continuous() so that
  it will handle type = "viridis" as the documentation states (@hlendway,
  #3079).

• scale_shape_identity() now works correctly with guide = "legend"
  (@malcolmbarrett, #3029)

• scale_continuous will now draw axis line even if the length of breaks is 0
  (@thomasp85, #3257)

• stat_bin() will now error when the number of bins exceeds 1e6 to avoid
  accidentally freezing the user session (@thomasp85).

• sec_axis() now places ticks accurately when using nonlinear transformations (@dpseidel, #2978).

• facet_wrap() and facet_grid() now automatically remove NULL from facet
  specs, and accept empty specs (@yutannihilation, #3070, #2986).

• stat_bin() now handles data with only one unique value (@yutannihilation
  #3047).

• sec_axis() now accepts functions as well as formulas (@yutannihilation, #3031).

• New theme elements allowing different ticks lengths for each axis. For instance,
  this can be used to have inwards ticks on the x-axis (axis.ticks.length.x) and
  outwards ticks on the y-axis (axis.ticks.length.y) (@pank, #2935).

• The arguments of Stat*$compute_layer() and Position*$compute_layer() are
  now renamed to always match the ones of Stat$compute_layer() and
  Position$compute_layer() (@yutannihilation, #3202).

• geom_*() and stat_*() now accepts purrr-style lambda notation
  (@yutannihilation, #3138).

• geom_tile() and geom_rect() now draw rectangles without notches at the
  corners. The style of the corner can be controlled by linejoin parameters
  (@yutannihilation, #3050).

ggplot2 3.1.1

This is a minor release with internal code changes to support the coming changes in the grid unit specification. This release includes no new features and no breaking changes

ggplot 3.1.0

ggplot2 3.1.0

Breaking changes

This is a minor release and breaking changes have been kept to a minimum. End users of ggplot2 are unlikely to encounter any issues. However, there are a few items that developers of ggplot2 extensions should be aware of. For additional details, see also the discussion accompanying issue #2890.

• In non-user-facing internal code (specifically in the aes() function and in
  the aesthetics argument of scale functions), ggplot2 now always uses the British
  spelling for aesthetics containing the word "colour". When users specify a "color"
  aesthetic it is automatically renamed to "colour". This renaming is also applied
  to non-standard aesthetics that contain the word "color". For example, "point_color"
  is renamed to "point_colour". This convention makes it easier to support both
  British and American spelling for novel, non-standard aesthetics, but it may require
  some adjustment for packages that have previously introduced non-standard color
  aesthetics using American spelling. A new function standardise_aes_names() is
  provided in case extension writers need to perform this renaming in their own code
  (@clauswilke, #2649).

• Functions that generate other functions (closures) now force the arguments that are
  used from the generated functions, to avoid hard-to-catch errors. This may affect
  some users of manual scales (such as scale_colour_manual(), scale_fill_manual(),
  etc.) who depend on incorrect behavior (@krlmlr, #2807).

• Coord objects now have a function backtransform_range() that returns the
  panel range in data coordinates. This change may affect developers of custom coords,
  who now should implement this function. It may also affect developers of custom
  geoms that use the range() function. In some applications, backtransform_range()
  may be more appropriate (@clauswilke, #2821).

New features

• coord_sf() has much improved customization of axis tick labels. Labels can now
  be set manually, and there are two new parameters, label_graticule and
  label_axes, that can be used to specify which graticules to label on which side
  of the plot (@clauswilke, #2846, #2857, #2881).

• Two new geoms geom_sf_label() and geom_sf_text() can draw labels and text
  on sf objects. Under the hood, a new stat_sf_coordinates() calculates the
  x and y coordinates from the coordinates of the sf geometries. You can customize
  the calculation method via fun.geometry argument (@yutannihilation, #2761).

Minor improvements and fixes

• benchplot() now uses tidy evaluation (@dpseidel, #2699).

• The error message in compute_aesthetics() now only provides the names of
  aesthetics with mismatched lengths, rather than all aesthetics (@karawoo,
  #2853).

• For faceted plots, data is no longer internally reordered. This makes it
  safer to feed data columns into aes() or into parameters of geoms or
  stats. However, doing so remains discouraged (@clauswilke, #2694).

• coord_sf() now also understands the clip argument, just like the other
  coords (@clauswilke, #2938).

• fortify() now displays a more informative error message for
  grouped_df() objects when dplyr is not installed (@jimhester, #2822).

• All geom_*() now display an informative error message when required
  aesthetics are missing (@dpseidel, #2637 and #2706).

• geom_boxplot() now understands the width parameter even when used with
  a non-standard stat, such as stat_identity() (@clauswilke, #2893).

• geom_hex() now understands the size and linetype aesthetics
  (@mikmart, #2488).

• geom_hline(), geom_vline(), and geom_abline() now work properly
  with coord_trans() (@clauswilke, #2149, #2812).

• geom_text(..., parse = TRUE) now correctly renders the expected number of
  items instead of silently dropping items that are empty expressions, e.g.
  the empty string "". If an expression spans multiple lines, we take just
  the first line and drop the rest. This same issue is also fixed for
  geom_label() and the axis labels for geom_sf() (@slowkow, #2867).

• geom_sf() now respects lineend, linejoin, and linemitre parameters
  for lines and polygons (@alistaire47, #2826).

• ggsave() now exits without creating a new graphics device if previously
  none was open (@clauswilke, #2363).

• labs() now has named arguments title, subtitle, caption, and tag.
  Also, labs() now accepts tidyeval (@yutannihilation, #2669).

• position_nudge() is now more robust and nudges only in the direction
  requested. This enables, for example, the horizontal nudging of boxplots
  (@clauswilke, #2733).

• sec_axis() and dup_axis() now return appropriate breaks for the secondary
  axis when applied to log transformed scales (@dpseidel, #2729).

• sec_axis() now works as expected when used in combination with tidy eval
  (@dpseidel, #2788).

• scale_*_date(), scale_*_time() and scale_*_datetime() can now display
  a secondary axis that is a one-to-one transformation of the primary axis,
  implemented using the sec.axis argument to the scale constructor
  (@dpseidel, #2244).

• stat_contour(), stat_density2d(), stat_bin2d(), stat_binhex()
  now calculate normalized statistics including nlevel, ndensity, and
  ncount. Also, stat_density() now includes the calculated statistic
  nlevel, an alias for scaled, to better match the syntax of stat_bin()
  (@bjreisman, #2679).

ggplot2 3.0.0

Breaking changes

• ggplot2 now supports/uses tidy evaluation (as described below). This is a
  major change and breaks a number of packages; we made this breaking change
  because it is important to make ggplot2 more programmable, and to be more
  consistent with the rest of the tidyverse. The best general (and detailed)
  introduction to tidy evaluation can be found in the meta programming
  chapters in Advanced R.

  The primary developer facing change is that aes() now contains
  quosures (expression + environment pairs) rather than symbols, and you'll
  need to take a different approach to extracting the information you need.
  A common symptom of this change are errors "undefined columns selected" or
  "invalid 'type' (list) of argument" (#2610). As in the previous version,
  constants (like aes(x = 1) or aes(colour = "smoothed")) are stored
  as is.

  In this version of ggplot2, if you need to describe a mapping in a string,
  use quo_name() (to generate single-line strings; longer expressions may
  be abbreviated) or quo_text() (to generate non-abbreviated strings that
  may span multiple lines). If you do need to extract the value of a variable
  instead use rlang::eval_tidy(). You may want to condition on
  (packageVersion("ggplot2") <= "2.2.1") so that your code can work with
  both released and development versions of ggplot2.

  We recognise that this is a big change and if you're not already familiar
  with rlang, there's a lot to learn. If you are stuck, or need any help,
  please reach out on https://community.rstudio.com.

• Error: Column y must be a 1d atomic vector or a list

  Internally, ggplot2 now uses as.data.frame(tibble::as_tibble(x)) to
  convert a list into a data frame. This improves ggplot2's support for
  list-columns (needed for sf support), at a small cost: you can no longer
  use matrix-columns. Note that unlike tibble we still allow column vectors
  such as returned by base::scale() because of their widespread use.

• Error: More than one expression parsed

  Previously aes_string(x = c("a", "b", "c")) silently returned
  aes(x = a). Now this is a clear error.

• Error: data must be uniquely named but has duplicate columns

  If layer data contains columns with identical names an error will be
  thrown. In earlier versions the first occuring column was chosen silently,
  potentially masking that the wrong data was chosen.

• Error: Aesthetics must be either length 1 or the same as the data

  Layers are stricter about the columns they will combine into a single
  data frame. Each aesthetic now must be either the same length as the data
  frame or a single value. This makes silent recycling errors much less likely.

• Error: coord_* doesn't support free scales

  Free scales only work with selected coordinate systems; previously you'd
  get an incorrect plot.

• Error in f(...) : unused argument (range = c(0, 1))

  This is because the oob argument to scale has been set to a function
  that only takes a single argument; it needs to take two arguments
  (x, and range).

• Error: unused argument (output)

  The function guide_train() now has an optional parameter aesthetic
  that allows you to override the aesthetic setting in the scale.
  To make your code work with the both released and development versions of
  ggplot2 appropriate, add aesthetic = NULL to the guide_train() method
  signature.

  # old
  guide_train.legend <- function(guide, scale) {...}
  
  # new 
  guide_train.legend <- function(guide, scale, aesthetic = NULL) {...}

  Then, inside the function, replace scale$aesthetics[1],
  aesthetic %||% scale$aesthetics[1]. (The %||% operator is defined in the
  rlang package).

  # old
  setNames(list(scale$map(breaks)), scale$aesthetics[1])
  
  # new
  setNames(list(scale$map(breaks)), aesthetic %||% scale$aesthetics[1])

• The long-deprecated subset argument to layer() has been removed.

Tidy evaluation

• aes() now supports quasiquotation so that you can use !!, !!!,
  and :=. This replaces aes_() and aes_string() which are now
  soft-deprecated (but will remain around for a long time).

• facet_wrap() and facet_grid() now support vars() inputs. Like
  dplyr::vars(), this helper quotes its inputs and supports
  quasiquotation. For instance, you can now supply faceting variables
  like this: facet_wrap(vars(am, cyl)) instead of
  facet_wrap(~am + cyl). Note that the formula interface is not going
  away and will not be deprecated. vars() is simply meant to make it
  easier to create functions around facet_wrap() and facet_grid().

  The first two arguments of facet_grid() become rows and cols
  and now support vars() inputs. Note however that we took special
  care to ensure complete backward compatibility. With this change
  facet_grid(vars(cyl), vars(am, vs)) is equivalent to
  facet_grid(cyl ~ am + vs), and facet_grid(cols = vars(am, vs)) is
  equivalent to facet_grid(. ~ am + vs).

  One nice aspect of the new interface is that you can now easily
  supply names: facet_grid(vars(Cylinder = cyl), labeller = label_both) will give nice label titles to the facets. Of course,
  those names can be unquoted with the usual tidy eval syntax.

sf

• ggplot2 now has full support for sf with geom_sf() and coord_sf():

  nc <- sf::st_read(system.file("shape/nc.shp", package = "sf"), quiet = TRUE)
  ggplot(nc) +
    geom_sf(aes(fill = AREA))

  It supports all simple features, automatically aligns CRS across layers, sets
  up the correct aspect ratio, and draws a graticule.

New features

• ggplot2 now works on R 3.1 onwards, and uses the
  vdiffr package for visual testing.

• In most cases, accidentally using %>% instead of + will generate an
  informative error (#2400).

• New syntax for calculated aesthetics. Instead of using aes(y = ..count..)
  you can (and should!) use aes(y = stat(count)). stat() is a real function
  with documentation which hopefully will make this part of ggplot2 less
  confusing (#2059).

  stat() is particularly nice for more complex calculations because you
  only need to specify it once: aes(y = stat(count / max(count))),
  rather than aes(y = ..count.. / max(..count..))

• New tag label for adding identification tags to plots, typically used for
  labelling a subplot with a letter. Add a tag with labs(tag = "A"), style it
  with the plot.tag theme element, and control position with the
  plot.tag.position theme setting (@thomasp85).

Layers: geoms, stats, and position adjustments

• geom_segment() and geom_curve() have a new arrow.fill parameter which
  allows you to specify a separate fill colour for closed arrowheads
  (@hrbrmstr and @clauswilke, #2375).

• geom_point() and friends can now take shapes as strings instead of integers,
  e.g. geom_point(shape = "diamond") (@daniel-barnett, #2075).

• position_dodge() gains a preserve argument that allows you to control
  whether the total width at each x value is preserved (the current
  default), or ensure that the width of a single element is preserved
  (what many people want) (#1935).

• New position_dodge2() provides enhanced dodging for boxplots. Compared to
  position_dodge(), position_dodge2() compares xmin and xmax values
  to determine which elements overlap, and spreads overlapping elements evenly
  within the region of overlap. position_dodge2() is now the default position
  adjustment for geom_boxplot(), because it handles varwidth = TRUE, and
  will be considered for other geoms in the future.

  The padding parameter adds a small amount of padding between elements
  (@karawoo, #2143) and a reverse parameter allows you to reverse the order
  of placement (@karawoo, #2171).

• New stat_qq_line() makes it easy to add a simple line to a Q-Q plot, which
  makes it easier to judge the fit of the theoretical distribution
  (@nicksolomon).

Scales and guides

• Improved support for mapping date/time variables to alpha, size, colour,
  and fill aesthetics, including date_breaks and date_labels arguments
  (@karawoo, #1526), and new scale_alpha() variants (@karawoo, #1526).

• Improved support for ordered factors. Ordered factors throw a warning when
  mapped to shape (unordered factors do not), and do not throw warnings when
  mapped to size or alpha (unordered factors do). Viridis is used as the
  default colour and fill scale for ordered factors (@karawoo, #1526).

• The expand argument of scale_*_continuous() and scale_*_discrete()
  now accepts separate expansion values for the lower and upper range
  limits. The expansion limits can be specified using the convenience
  function expand_scale().

  Separate expansion limits may be useful for bar charts, e.g. if one
  wants the bottom of the bars to be flush with the x axis but still
  leave some (automatically calculated amount of) space above them:

  ggplot(mtcars) +
      geom_bar(aes(x = factor(cyl))) +
      scale_y_continuous(expand = expand_scale(mult = c(0, .1)))

  It can also be useful for line charts, e.g. for counts over time,
  where one wants to have a ’hard’ lower limit of y = 0 but leave the
  upper limit unspecified (and perhaps differing between panels), with
  some extra space above the highest point on the line (with symmetrical
  limits, ...

ggplot2 2.2.1

• Fix usage of structure(NULL) for R-devel compatibility (#1968).