v0.11.1

Breaking changes

• The extract_body() function now, by default, will not display columns that have been hidden (e.g., by cols_hide() or cols_merge*()); the previous behavior can be restored by using incl_hidden_cols = TRUE.

• PDF table output has changed from a longtable environment to a floating environment (using tabular*); use tab_options(latex.use_longtable = TRUE) to restore the LaTeX-table output behavior from previous releases.

Improvements to the LaTeX output format

• PDF output now defaults to a full-width floating environment using tabular*. Float position can be controlled by the latex.tbl.pos argument in tab_options. Quarto users can alternatively use the tbl-pos argument to control positioning. To use a longtable environment instead, use tab_options(latex.use_longtable = TRUE). (@AronGullickson, #1588)

• Creating a caption with tab_caption() will now be preserved in LaTeX output with as_latex(). Cross-referencing a table using the internal cross-referencing system of bookdown is now enabled for PDF and HTML outputs (for HTML, set options("htmltools.preserve.raw" = FALSE)). Quarto users should use the tbl-cap and label cell options. (@nielsbock, #1800)

• Improved the centering of the stubhead label in Latex when row_group_as_column = TRUE and the width of the row name column is specified. (@kbrevoort, #1804)

• LaTeX now correctly renders if text size is changed (#1885). (@olivroy, #1899)

• Fixed a bug that caused an error in LaTeX when (1) row_group_as_column = TRUE, (2) the row groups were specified using tab_row_group(), and (3) the user specified a width for the row label column. (@kbrevoort, #1804)

• Fixed an issue where column widths weren't set properly using col_widths() for LaTeX output (#1837). (@snhansen, #1867)

Interactive table support

• Interactive tables will show no border if opt_table_lines(extent = "none") is specified (#1307).

• Interactive tables now respect more styling options, namely: column_labels.background.color, row_group.background.color, row_group.font.weight, table_body.hlines.style, table.font.weight, table.font.size, and stub.font.weight. (#1693)

• opt_interactive() now works when columns are merged with cols_merge(). (@olivroy, #1785)

• opt_interactive() now works when columns are substituted with sub_*(). (@olivroy, #1759)

• More support for cells_stubhead() styling and footnotes in interactive tables.

Minor improvements and bug fixes

• New arguments have been added to extract_body() to better control which columns will be present in the returned data (#1875). (#1889)

• The locale argument of gt() now defaults to getOption("gt.locale") if set. (#1894)

• The optional argument levels was added to cells_column_spanners() to constrain targeting of column spanners by their level. (@obsaditelnost, #1858)

• Fixed a bug in using pct() column widths with as_gtable() (#1771). (@teunbrand, #1776)

• Fixed a bug where gt(row_group_as_column = TRUE) would create the wrong layout with as_gtable() when all groups are unique. (@olivroy, #1803)

• Showing currency symbols in grid output now works. (@olivroy, #1788)

• data_color() no longer errors when a tidyselect selection is empty (like fmt_*() functions) (#1665). (@olivroy, #1795)

• Fixed an issue when using text_case_match(.replace = "all", .locations = cells_column_spanners()). (@olivroy, #1823)

• Improved performance during footnote rendering. (@olivroy, #1818)

• Improved footnote rendering in Quarto with fmt_markdown() (#1773). (@olivroy, #1860)

• Fixed an issue where md("") would fail in Quarto (#1769). (@olivroy, #1772)

• Fixed a bug where spanners didn't render correctly in Quarto (#1839). (#1856)

• vec_fmt_markdown() works correctly inside Quarto again (#1840). (@olivroy, #1841)

• Fixed an issue where tab_spanner_delim() would fail to resolve a duplicate ID value. (@olivroy, #1821)

• Fixed an issue with multiple text_replace() calls would produce bad results with cells_column_labels(). (@olivroy, #1824)

• cols_add() works correctly in more cases. (#1893)

• tab_footnote() now correctly adds footnote marks in the cells_stub_summary() and cells_stub_grand_summary() locations (#1832). (@olivroy, #1833)

• tidyselect::where(), tidyselect::all_of(), and tidyselect::any_of() are now re-exported by gt. (#1830)

• Fixed an issue where md() and fmt_markdown() would render factors as their numeric levels rather than their text labels (#1882). (@rossellhayes, #1883)

• vec_fmt_*() (and incidentally cols_nanoplot()) should be faster now. (@olivroy, #1888, #1891, #1896, #1901)

• Improved error messages for the text_transform() function if locations couldn't be resolved. (@olivroy, #1774)

• tab_row_group() gives a more precise error message when rows can't be resolved correctly (#1535). (@olivroy, #1770)

v0.11.0

New features

• gt tables can now be rendered in the grid graphics system with the new as_gtable() function (#180, #420, #509, #961, #1652, #1701). Thank you @teunbrand for this valuable contribution! (#1563, #1648, #1655, #1727)

• With the new fmt_chem() function, it's possible to format chemical formulas and chemical equations in the table body. (#1636)

• We added fmt_email() we can transform email addresses in table cells into usable 'mailto:' links. This operates similarly to fmt_url() but removes some unneeded anchor tag attributes (#1616). (#1649)

• The new fmt_tf() function lets you format logical values in a gt table. You could express these as a 'yes' or 'no', a 'true' or 'false', or, perhaps use pairings of complementary symbols that make sense in a table (#1650). (#1654)

• The new fmt_country() was added to help with making country names consistent and localizable. The input text can either consist of 2- or 3-letter ISO 3166-1 country codes. (#1645, #1696)

• Math formulas (written in LaTeX) can now be rendered to HTML through md() and fmt_markdown(). The LaTeX formulas can be set between $ or $$ delimiters for inline and block rendering styles. While this requires the katex package to work (it's an optional dependency), the rendering of formulas is dependency-free in the output (#375, #616, #1163). (#1578)

• We added the unit_conversion() helper function to help you perform unit conversions. This function returns a conversion factor for transforming a value from one form of measurement units to a target form. This works wonderfully in the scale_by argument that is present in many number-based fmt_*() functions. (#1704, #1709, #1710, #1729)

• The plot() method was added for gt_tbl so that one can quickly create a reprex visual example (e.g., exibble |> gt() |> plot()). (#1742)

• info_time_style() and info_date_style() gain a locale argument to preview time and date formatting in a specific locale. (#1747)

• Several datasets were added to the package: reactions, photolysis, peeps, nuclides, gibraltar, and films. (#1619, #1641, #1664, #1668, #1686, #1695)

Improvements to the LaTeX output format

• A host of improvements were made to LaTeX output such that styling made through tab_options() and tab_style() is now honored. The styling implemented now includes: (1) font size, underlining, italicizing, bolding, indenting text, coloring text, filling cell backgrounds, and modifying text case (i.e., uppercase, lowercase, sentence case) (#1472). (#1594, #1603, #1746, thanks so much @kbrevoort)

• Text in spanner labels is now nicely wrapped in LaTeX output, where before there would be no wrapping at all (#1656). (#1716, #1746)

• The gt() setting of row_group_as_column == TRUE is now implemented in LaTeX output. (#1716)

• LaTeX output now understands and can convert from lengths expressed in pixels (an internal conversion to pt units is performed) (#1582). (#1595)

• A fix was made to correct an error related to converting px units to pt units for LaTeX output. (#1591, thank you @AaronGullickson)

Improvements to interactive HTML tables via opt_interactive()

• The performance of interactive table rendering through opt_interactive() has been improved (#1289). (#1735)

• opt_interactive() now shows row names if rownames_to_stub = TRUE (#1702). (@olivroy, #1706)

• There's now better support for displaying group columns, if present, in interactive tables (#1705). (@olivroy, #1725, #1758, #1760)

• Interactive tables can now display the stub header label created with tab_stubhead(). (@olivroy, #1758).

• There is no longer an error when having a hidden column under a spanner label in an interactive HTML table (i.e., opt_interactive()) context. (#1629)

• Interactive tables now render the first level of column groups added by tab_spanner() (#1618). Thanks @obsaditelnost for your work on this! (#1623)

• The opt_interactive() function now has the height argument to help specify the height of an interactive table (#1544). (@olivroy, #1723).

• When using opt_interactive(use_pagination = FALSE) the pagination controls are now entirely removed instead of merely being non-functional (#1542). (#1642)

• When supplying a locale value to an interactive table, we now ensure that the supplied locale is normalized (so "fr_CH" will be internally corrected to "fr-CH") (#1637). (#1643)

Minor improvements and bug fixes

• The fill_color argument in fmt_icon() now allows the use of named vector/list to apply different colors to different icons (#1560). (#1647)

• fmt_flag() gains a locale argument so that SVG title text displays country names in any locale language. (#1645)

• To better represent missing values in nanoplots, the "marker" option has been added in cols_nanoplot() (#1567). (#1587)

• Many of the info_*() functions were upgraded with better functionality and a refreshed look. (#1720, #1748, #1754)

• The metro dataset has been updated with six new Line 11 stations that opened on 2024-06-13. (#1708)

• For HTML tables produced in Quarto, we now use the data-qmd-base64 attribute along with base64 text when passing Markdown text to Quarto (which handles the Markdown conversion) (#1487, #1488). (#1688, #1690, #1734)

• Scientific and engineering notation will no longer unexpectedly wrap lines in HTML for smaller column widths. (#1621)

• Footnote marks in HTML table outputs no longer unexpectedly increase the text line height (#1556). (#1644)

• We now ensure tables print correctly when called from inline code in .qmd or .Rmd documents (#1055). (#1689)

• data_color() throws a more informative error if an internal calculation failed (e.g., presence of infinite values) (#1373). (@olivroy, #1585)

• data_color() provides an error message if rows doesn't resolve to anything (#1659). (@olivroy, #1660).

• summary_rows() now throws a more informative error message that you should use grand_summary_rows() if no row groups are detected (#1292). (@olivroy, #1752)

• Boxplot-based nanoplots now correctly determine which data points are outliers. (#1756)

• The issue where having a row group set as a column in the stub, corrupting the setting of column widths in HTML output tables, has been fixed (#1253, #1510). (#1744)

• The rendering of footnotes and source notes received fixes to ensure that all requested notes are inserted and also that the typesetting options in tab_options() is implemented (#1505, #1615). (#1620, #1751)

• Using a fmt_*() function while the stub is composed of Markdown or HTML no longer results in an error (#1600). (#1627)

• The combination of using cols_label() + summary_rows() / grand_summary_rows() no longer causes an error for RTF output through gtsave() (#1233). (#1596)

• We now ensure that id values are stripped of HTML tags when using tab_row_group(). Previously this condition would create corrupted row group labels (#1143). (#1750)

• cols_hide() no longer errors if a column wasn't supplied. Error messages are also clearer when supplying a column that doesn't exist (#1631). (#1632)

• cols_units(), tab_footnote() and tab_style() now give better error messages when locations is not correctly specified (#475). (@olivroy, #1638, #1640, #1733)

• Error messages are now improved for rm_spanners(), rm_footnotes(), and rm_source_notes() (#1638). (#1661)

• Error messages for the vec_*() functions are now standardized with check_vector_valid(). (#1662)

• Running testthat tests no longer fails on Windows due to gtsave() path incompatibilities (#1626). (#1627)

• Tests run on Windows are less noisy now (no more render related warnings). (#1634)

• Much code refactoring, linting, and test improvements were performed by @olivroy, resulting in many performance gains. (#1666, #1683, #1694, #1699, #1715, #1718, #1728, #1730, #1731, #1732, #1733, #1753)

• gt now depends on R 3.6. (#1599, #1731).

Documentation enhancements

• Many small documentation fixes were performed. (#1555, #1598, #1633, #1658, #1663, #1676, #1685, #1700, #1703, #1717, #1719, #1721)

• More examples were added to further demonstrate how various functions can be used. (#1646, #1667, #1677, #1698, #1724, #1737)

v0.10.1

Improvements to nanoplots

• Box plots can now be generated via cols_nanoplot() by using plot_type = "boxplot". These plots are laid out horizontally and will, by default, share the same plot axis across rows. (#1527)

• We can now have single line bar plots generated through cols_nanoplot(). If the plot_type is set to "bar" and single values are found, then horizontal bars will be generated and will be comparable across rows. (#1514, #1515, #1519)

• The autohide argument was added to the cols_nanoplot() function so that columns containing input data for nanoplots could be conveniently hidden from final presentation. (#1533)

• Added option (the data_area_fill_color arg in nanoplot_options()) to change fill color of nanoplot data area for line-type plots (#1521). (#1534)

Minor improvements and bug fixes

• The performance of rendering bigger tables as HTML has been improved and is now up to three times faster than before. (#1470, thanks @mgirlich)

• Introduced a small performance improvement by no longer calling utils::packageVersion() internally (#1524). (#1525, thank you @slodge)

• Code and test refactoring was performed to generally improve performance and code readability. (#1480, thanks @olivroy)

• The gtsave() function now returns the file path invisibly instead of TRUE. (#1478, thank you @olivroy)

• Most functions now produce better error messages if not provided with a gt_tbl object. (#1504, c/o @olivroy)

• The URL formatting through fmt_url() has been improved by preventing link text breaking across lines (#1509). (#1537)

• We now remove some unnecessary newlines in the HTML text produced by as_raw_html(), which caused an issue when integrating gt tables into blastula email messages (#1506). (#1520)

• The tab_spanner_delim() now lets you use delim strings longer than a single character (#1469). (#1513)

• Fix for footnotes in LaTeX tables where no footnote marks are to be added; this previously showed "NA" as the mark in the footer area but this is no longer displayed (#1416). (#1512, thanks @kbrevoort)

• LaTeX tables can now have their overall width specified (#119, #329). (#1495, thank you @kbrevoort)

• Fix issue where a cols_width() specification involving percentage values fails for LaTeX tables (#1465). (#1495, thanks again @kbrevoort!)

• Several documentation fixes were made to address inconsistencies and improve clarity. (#1491)

v0.10.0

Nanoplots

• We can now add in little plots called nanoplots to a gt table (#299, #515). (#1431, #1439, #1445, #1453, #1458, #1459, #1461, #1462)

• The function cols_nanoplot() adds a new column that contains the plots. The data can be obtained from one or more columns in the table. A helper function called nanoplot_options() allows for altering the composition and styling of the nanoplots in the new column.

• There are two basic types of nanoplots available: "line" and "bar". A line plot shows individual data points and has smooth connecting lines between them to allow for easier scanning of values. You can opt for straight-line connections between data points, or, no connections at all (it's up to you). The data you feed into a line plot can consist of a single vector of values (resulting in equally-spaced y values), or, you can supply two vectors representative of x and y.

• A bar plot is built a little bit differently. The focus is on evenly-spaced bars (requiring a single vector of values) that project from a zero line, clearly showing the difference between positive and negative values.

• By default, any type of nanoplot will have basic interactivity. One can hover over the data points and vertical guides will display values ascribed to each. A guide on the left-hand side of the plot area will display the minimal and maximal y values on hover.

Other great new features

• Brand new rows can be added to a gt table with the new rows_add() function. The user can supply the new row data through name value pairs. You have control over where they are placed by way of the .before and .after arguments (new rows are added to the bottom of the table by default). You can also add empty (i.e., all NA) rows with the .n_empty option (#698). (#1323)

• To complement rows_add(), the cols_add() function was added. New columns can indeed be added to a gt table with this function, which has an interface close to that of dplyr::mutate(). (#1367)

• You can now use an empty table as the starting point for a gt table. This can be used in conjunction with cols_add() and rows_add() to build a table piece-by-piece in specific workflows/settings. What constitutes empty tables can be any of: 0 x 0 tables, 0 x n tables (no rows, some columns), or n x 0 tables (some rows, no columns; treated the same as 0 x 0 tables). (#1376)

• There is now a way to better express measurement units and we do this in gt with something called units notation. With an intuitive and easy-to-learn syntax, gt will ensure that any measurement units are formatted correctly no matter what the output type is. We can format units in the table body with fmt_units(), we can attach units to column labels with cols_units(), and we can integrate units notation in the already-available cols_label() and tab_spanner() functions (#417, #533). (#1357, #1426, #1446)

• A very useful new helper function, from_column(), has been added so you can fetch values (for compatible arguments) from a column in the input table. For example, if you are using fmt_scientific() and the number of significant figures should vary across the values to be formatted, a column containing those values for the n_sigfig argument can be referenced by from_column(). (#1392, #1393, #1395, #1396, #1399, #1403)

• With the new fmt_icon() function we are able to add icons from the Font Awesome icon library. It works in a way that's similar to fmt_flag(), in that identifiers in the formatted cells are transformed in-place to SVG-based icons. (#1413)

• The info_icons() and info_flags() functions have been added to help people know about the valid codes for flags and for icons (when using fmt_icon() and fmt_flag()). (#1421)

• We added the extract_body() function, which lets you pull out a data frame associated with the body cells. Importantly, this extraction can happen at different stages of the table build (e.g., "init", "text_transformed", etc.), allowing a user to have access to a table of formatted body cells for different applications (such as verification of formatting, debugging new formatting functions, etc.) (#1441). (#1449)

• Interactive HTML tables (usually generated through use of opt_interactive()) can now use localized labels/controls. So when using any of the 574 supported locales in gt, an interactive table will be fully translated to the language of the locale (#1308). (#1389)

• The illness and constants datasets were added. Both datasets have a units column and this is useful for making examples with the fmt_units() function. (#1357)

Improvements to the Word output format

• Processing to Word output now escapes HTML in more places (#1378). (#1303)

• The Word output format now uses the side argument present in summary_rows() and grand_summary_rows() to place the new summary rows either the top or bottom of the row group (with summary_rows()) or table as a whole (with grand_summary_rows()). (#1325)

• Tables rendered as Word output can now handle the specific case where a table with summary rows doesn't have row names. (#1325)

• Summary rows in Word output tables can now be placed at the top or bottom of a group (or at the top or bottom of the table). (#1402)

• Word output tables can now contain images. This entails compatibility with the fmt_image() function, and, images (local and remote) can be inserted through Markdown (#1272). (#1273)

Documentation enhancements

• The gt website has been updated with a slightly different look; section names have been updated for consistency (#1419). (#1287, #1340, #1341, #1444)

• We've improved the formatting of arguments in the documentation so that they all have short titles and descriptions regarding expected inputs and default values. This looks great both in the internal R help pages and in the pkgdown-generated website (#1290). (#1338)

• Several small documentation updates were made, with an emphasis on improving examples (#1304, #1349, #1369). (#1293, #1316, #1324, #1329, #1330, #1331, #1334, #1381, #1383, #1395, #1404, #1442, #1454)

Minor improvements and bug fixes

• The tab_spanner_delim() function was given a limit argument so that splitting from a particular side can stop early and precisely. (#1328)

• The width argument was added to fmt_image(). With this change you can variously set the width, the height, or both.

• Significant figures support has been added to the fmt_scientific() and vec_fmt_scientific() functions; there is a new n_sigfig argument in both. (#1411)

• The cols_merge_range() function now has a locale argument. Range patterns across locales are different (can involve the use of a single hyphen, en dash, em dash, tilde, etc.) and so it does make sense to follow the convention of a locale if provided (#158). (#1423)

• The fmt_url() function now has a few more options for adding anchor tag attributes ("target", "rel", "referrerpolicy", and "hreflang"). Thanks @elipousson for the work on this! (#1428). (#1452)

• We now have rudimentary support for defining column widths for LaTeX output tables (with cols_width()). This accepts length values in 'px' which and automatic conversion to 'pt' values is performed to maximize compatibility with different LaTeX flavors (#634, #851, #1417). (#1371, #1450)

• It's now possible to use background fill colors and perform text coloring and emboldened/italicized text within the body cells of LaTeX tables. This is commonly performed through the use of tab_style() and data_color() (#84, #869). (#1352)

• The gtsave() function now works with gt_group objects (usually generated through gt_split() or gt_group()) (#1354). (#1365)

• All gt_group objects can now be printed using R Markdown or Quarto (#1286). (#1332)

• When using fmt_currency() with a locale value set, gt will now use that to automatically select the locale's default currency. While some countries can have multiple currencies, we opt for the most-widely used currency (users could alternatively specify the currency code and info_currencies() contains all supported currencies used in the package) (#1346). (#1347)

• The columns argument in cols_hide() and cols_unhide() can now accept NULL (i.e., no columns resolved). These functions will no longer error in such a case (#1342). (#1343)

• The countrypops dataset was updated with recent (as of August 2023) World Bank data that revises population estimates and brings the final year up to 2022. All examples, tests, and articles using the dataset were also updated. (#1410)

• A few refinements were made to some of the system font stacks defined in system_fonts(). (#1447)

• The Databricks notebook environment is now detected by gt, so tables will now be automatically displayed without having to call extra printing functions. (#1427)

• Display issues with scientific and engineering notation formatting (lack of dropping trailing zeros, as promised with the drop_trailing_zeros argument) were fixed. Part of the fix involves adding the drop_trailing_dec_mark argument to the four functions fmt_scientific(), fmt_engineering(), vec_fmt_scientific() and vec_fmt_engineering(). (#1380)

• Fixed an incorrect country code reference for the Netherlands that would cause an incorrect flag to appear when using fmt_flag(). (#1319)

• Many new flags were added to fmt_flag() (#1333, #1335). (#1336)

• In some cases, there was incorrect rounding of duration values when using fmt_duration(). This is now fixed, thanks to @rcannood (#1374). (#1375)

• Fixed an issue with cols_label_with() where column names wouldn't be relabeled if the resolved columns were only a subset of the total columns available. (#1326)

• Fixed a LaTeX bug where some characters f...

v0.9.0

New features

• The summary_rows() and grand_summary_rows() functions have been rewritten extensively to allow more flexibility in providing aggregation information to fns. The documentation for both functions will walk you through all the new ways you can express your aggregations. The formatting of cells in new summary rows is now easier thanks to the new fmt argument. We can provide a single formatting expression or a number of them in a list. The dedicated documentation section for formatting can walk you through all of this. Furthermore, we now have the option to place summary rows at top or bottom of a group (in summary_rows()) or table (in grand_summary_rows()) with the new side argument (#177, #270, #292, #545, #581, #784, #785, #840). (#1159, #1171)

• The data_color() function has been completely overhauled to make it easier to use and also more powerful. It gains a rows argument to better constrain where the colorization should take place. Instead of having to use functions from scales to define the coloring scheme, you can now use the method argument to handle most cases (otherwise, the new fn argument allows for a color-mapping function to be used). The new palette argument allows for flexible input of color palettes. One or more adjacent columns can receive the coloring treatment by way of the new target_columns argument. Coloring can now also be applied in a row-wise fashion by using direction = "row". Despite all of these changes, older code should still work with the revamped function (#116, #633, #637, #1103, #1119). (#1147)

• We now have a much improved cols_merge() function! It has gained a rows argument for limiting which body cells get merged together. The pattern argument now has a more advanced syntax. It now has the following delimiters and rules: (1) { } for arranging single column values in a row-wise fashion, and (2) << >> to surround spans of text that will be removed if any of the contained { } yields a missing value. The new << >> delimiters will help in cases where NA values are present and should be omitted during the merging process (#1167). (#1144, #1168)

• The cols_label() function has been improved by allowing formula expressions to be given to .... With two-sided formulas (e.g., <LHS> ~ <RHS>), the left-hand side corresponds to selections of columns and the right-hand side evaluates to single-length values for the label to apply. Named arguments in the format of <column name> = <label> are still also valid as input for simple mappings of column name to label text. Also, there's a .fn argument, which takes a function to be applied to all provided label values (#1146). (#1255, #1256, #1265)

• Interactive HTML tables can now be generated in gt. Options for enabling interactive table rendering are available in tab_options() and also in opt_interactive(). The collection of use_* arguments in opt_interactive() allow for control of pagination, global search, filtering, sorting, and more. The documentation within opt_interactive() demonstrates how to generate HTML tables with interactive elements (#207). (#902, #1252, #1261, #1278, #1280)

• A new function, cols_label_with(), has been added for relabeling columns via functions. It allows for such functions to be applied to arbitrary sets of columns, taking advantage of tidyselect. It's like dplyr::rename_with() for your gt table, but it doesn't need unique names for labeling (#701). (#1158, thank you @christopherkenny!)

• We added fmt_spelled_num() and vec_fmt_spelled_num() so that numeric values could be transformed to spelled out numbers. Any values from 0 to 100 can be spelled out according to a given locale value. For example, the value 23 will be rendered as "twenty-three" if the locale is an English-language one (or, not provided at all); should a Swedish locale be provided (e.g., "sv"), the output will instead be "tjugotre". (#1198)

• The fmt_index() and vec_fmt_index() functions were added in so that numeric values could undergo transformation to index values, which are usually based on letters (depending on the locale chosen). For example, the value 5 will be rendered as "E" if the locale is an English-language one. The characters chosen for indexing here are based on character sets intended for ordering (often leaving out characters with diacritical marks). (#1189)

• Should body cells contain URLs, the new fmt_url() function can be used to make them navigable links. There are several options provided that control how the links should be styled (conventional underline style or button-like text) (#609). (#1205, #1220, #1221)

• To more easily insert graphics into body cells, we have added the fmt_image() function. This allows for one or more images to be placed in the targeted cells. The cells need to contain some reference to an image file, either: (1) complete http/https or local paths to the files; (2) the file names, where a common path can be provided via path; or (3) a fragment of the file name, where the file_pattern argument helps to compose the entire file name and the path argument provides the path information. (#1240)

• The new fmt_flag() function helps with inserting a flag icon (or multiple) in body cells. The input cells need to contain 2-letter ISO 3166-1 country codes (e.g., Mauritius with the "MU" country code). This function will parse the targeted body cells insert the appropriate flag icon. Multiple flags can be included per body cell by separating country codes with commas. (#1241, #1244, #1251)

• With fmt_bins(), you can format bin or interval syntax (returned from cut()) to a form that presents better in a display table. It's possible to format the values of the intervals with the fmt argument, and, the separator can be modified with the sep argument. (#1250)

• The new fmt_auto() function has been added to automatically apply formatting of various types in a way that best suits the data table provided. For example, the function will attempt to format numbers such that they are condensed to an optimal width, and, currency values can be detected (by currency codes embedded in the column name) and formatted in the correct way (#248). (#1176)

• Three text_*() functions were added to better enable transformations of formatted text in various table locations: text_replace(), text_case_when(), and text_case_match(). (#1238)

• We can use any of 15 system font stacks available in gt! These are families of themed system fonts that are widely available in different computer systems. Examples include 'System UI' (the default font stack), 'Humanist', and 'Neo-Grotesque'. The new helper function system_fonts() can be used to access the various font stack and is useful when specifying font values in the cell_text() function. The opt_table_font() function has been upgraded to include a stack argument (which accepts keywords for the different font stacks). (#1275)

• We now have a new family of functions that let you deal with a multitude of gt tables, all at once. You can create a gt_group object with multiple tables via the gt_group() function, or, split a single table into multiple with gt_split(). There are some advantages to having a group of tables bundled together like this. You could apply options that pertain to all tables yet still access the individual tables to give them their own specialized modifications. They all print together at once too! For HTML, each table will be separated by a line break whereas in paginated formats (e.g., RTF, Word, etc.) the tables are separated by page breaks. (#1216, #1219, #1226, #1248, #1249, #1262, #1264)

• There is now a choice of which Markdown engine to use for Markdown rendering (in fmt_markdown() and vec_fmt_markdown()) and a default default choice. The markdown package can be chosen instead of commonmark to perform conversions of Markdown to HTML and LaTeX (and it's now the default). One of many benefits to this change is the parsing of subscripts (using ~x~) and superscripts (using ^x^) (#725). (#1254)

• Added the footnote formatting options footnotes.spec_ref and footnotes.spec_ftr in tab_options() and exposed these options further in a new function called opt_footnote_spec(). Specifications for footnote formatting can be produced for the footnote marks that alight to the targeted text within cells from various locations in the table, or, for the footnote marks that appear in the table footer. (#1268, #1271)

• The towny dataset was added. It provides a fun, population-based dataset with place names, URLs, coordinate information, and population figures with repetition (values for different census years). (#1184, #1199)

• The metro dataset is new. It has 308 rows, where each corresponds to a different Paris Metro station (#1231).

• The rx_adsl and rx_addv datasets have been added. These clinical trial toy datasets are both featured in a new vignette that explains how gt can be used for table generation in the Pharma space (#1145, #1200; thank you @alex-lauer!).

Documentation enhancements

• Addition of a new vignette (Case Study: Clinical Tables) that uses the rx_adsl and rx_addv datasets. (#1145, #1217, #1224; all contributed by @alex-lauer)

• All examples in the documentation were modernized by using the |> instead of the %>%. A lot of explanatory text was improved through rewriting. The ordering of documentation sections was improved. And functions that had very little documentation (like text_transform()) were greatly improved (#1087). (#1166, #1185, #1192)

• Several small documentation and housekeeping updates. (#1133, #1134, #1135, #1136, #1190, #1206, #1218, #1242, #1246, #1258, #1260, #1266, #1270, #1276, #1281)

• Added a citation file for the package. (#1128, #127...

v0.8.0

New features

• The new function sub_values() is here for subbing in arbitrary text in body cells, based on value, a regex, or a specialized function. It's like a supercharged find-and-replace for your gt table (#703). (#1066)

• We want the setting of styles to be as easy as possible so, to that end, we've added a helpful new function: tab_style_body(). This function is a bit like sub_values() and a bit like tab_style(). The idea is that basic style attributes can be set based on values in the table body. We can target body cells though value, regex, and custom matching rules, and, apply styles to them and their surrounding context (e.g., styling an entire row or column wherein the match is found). The help file at ?tab_style_body provides copious examples w/ eye catching visuals (#863). (#1108)

• The fmt_date(), fmt_time(), and fmt_datetime() functions have been improved so much, I don't even know where to begin. The fmt_datetime() function has an enhanced format argument that accepts CLDR-style formatting patterns (much more powerful and flexible than the strptime() patterns). The fmt_datetime() function will detect the pattern type given in format. The date_style and time_style arguments still exist but have been extended to accept more style keywords. There is now the concept of flexible date and time formats that are locale-aware (the locale argument is new to these functions). The info_date_style() and info_time_style() functions have been enhanced to give you display tables will all the style possibilities. All of this goodness extends to the vec_fmt_*() variants (vec_fmt_date(), vec_fmt_time(), and vec_fmt_datetime()). The documentation for all of this has been overhauled and provides many explanations and examples for your edification (#531, #1052). (#1053)

• We can now have decimal alignment for numeric values and this made possible with the new cols_align_decimal() function. The function ensures that columns targeted are right-aligned, that accounting notation is supported, and that footnote marks don't interfere (#139, #144, #560, #673). (#1058, #1092; assist by the inimitable @steveputman!)

• We can easily add a caption to a gt table (or replace an existing one) with the new and convenient tab_caption() function (#1048). (#1051, thank you @billdenney!)

• So you've got a gt table and you want a value or two from that table to appear elsewhere in your R Markdown or Quarto document. Previously, you had to recreate it in some way but now you can use extract_cells()! That function will extract a vector of cell data from a gt_tbl object. The output vector will have the cell data formatted in the same way as the table. (#1045, #1047)

• The as_raw_html() function is useful for generating an HTML string for table-in-HTML-email situations and for HTML embedding purposes (with that inline_css = TRUE option). While the CSS-inlining has been mostly fine, it had two major problems: (1) it was slow, and (2) the underlying R code was so underpowered that it just couldn't keep up with changes to our SCSS styles. This is now solved by integrating a package that uses the juice JS library (we call it juicyjuice!). This solution is far more performant and correct (#455, #540, #837, #858, #915, #1093). (#1114)

• There's now padding around an HTML table! And you can even control the left/right (container.padding.x) and top/bottom (container.padding.y) padding values through tab_options()! This is very helpful since tables were way too close to paragraphs of text in rendered HTML documents produced by R Markdown and Quarto (#590, #1105). (#1116)

• The table stub can now freely merge with other columns using any of the cols_merge*() collection of functions. This is great if you want to independently format the stub and other columns and then bring them together in interesting ways. (#1122)

• We've added the ability to treat columns containing row names or group names as Markdown text. This provides the process_md argument in the gt() function. If this option is taken (example below) then the supplied names will be treated to ensure that there are unique IDs for later targeting (for tab_style(), tab_footnote(), etc.) (#694). (#1097)

  dplyr::tibble(x = c("**a**", "b", "*c*"), y = c("*A*", "*B*", "*B*"), z = 1:3) %>%
    gt(rowname_col = "x", groupname_col = "y", process_md = TRUE)

• If you find yourself not knowing the ID values of certain cells in the table (sometimes necessary for adding footnotes, styles, etc.) the new tab_info() function can help! Use that with a gt object and you'll get an informative table that summarizes all of the table's ID values and their associated labels (#1102). (#1104)

• Much of gt is about adding things to a table but what about doing the opposite (taking things away)? The new family of rm_*() functions (rm_header(), rm_stubhead(), rm_spanners(), rm_footnotes(), rm_source_notes(), and rm_caption()) let us safely remove parts of a gt table. This can be advantageous in those instances where one might obtain a gt table though other means (like from another pkg) but prefer to excise some parts of it (#790, #1005). (#1073)

• The table stub column can now have its contents formatted with any of the fmt_*() functions. We also added the stub() helper function to help select the stub column (#747). (#1038)

• The new formatter function fmt_roman() lets us easily format numbers to Roman numerals (either as uppercase or lowercase letters). The vec_fmt_roman() vector-formatting function was also introduced here. (#1035)

• The data_color() function allows us to color the background of cells based on data, and gt smartly chooses a text color that provides the most contrast between text and background. We wanted to improve that feature so now data_color() has a contrast_algo argument that allows us to choose between two color contrast algorithms: "apca" (Accessible Perceptual Contrast Algorithm; new, and the default now) and "wcag" (Web Content Accessibility Guidelines). (#1062)

• The accessibility of structurally-complicated gt tables (i.e., multi-level headings, irregular headers, row groups, etc.) was addressed by work that follows the W3C WAI (Web Accessibility Initiative) guidance. Now, screen readers can better describe gt tables with such complex structures (#926). (#988, thanks @jooyoungseo!)

Minor improvements and bug fixes

• The docx output format is now better detected in R Markdown and Quarto (#1040). (#1084, thanks @cderv!)

• Replaced all match.arg() calls with rlang's match_arg() for better error output (#672). (#1099, thanks @mojister!)

• Project website improvements; we now have a doublet of sites: (1) https://gt.rstudio.com and (2) https://gt.rstudio.com/dev (#1074, thanks @ddsjoberg!)

• The dependency on the stringr package has been removed by replacing the remaining stringr-based function calls in the package with base-R equivalents. (#1043)

v0.7.0

New features

• We can now export gt tables as Word documents. This is thanks to @thebioengineer (Ellis Hughes!) who not only made this type of output work through gtsave() (writes the .docx file) but also through as_word() (gives you an OOXML string) (#121, #929). (#962, #986, #1016)

• A whole new family of vector formatting functions (of the form vec_fmt_*()) has entered the package! After all, why should columns have all the formatting fun? If you have vectors in need of formatting, we have 14 of these functions, adapted from their corresponding fmt_*() functions. You can control the output style of the resulting character vector with the output argument. This can either be "auto" (the default), "plain", "html", "latex", "rtf", or "word". In knitr rendering (i.e., Quarto or R Markdown), the "auto" option will choose the correct output value; this is great for incorporating formatted values in inline text. (#899, #995)

• It's now incredibly easy to have colorful and stylish tables thanks to the new opt_stylize() function. We provide you with a carefully curated set of background colors, line colors, and line styles. There are six styles and six color variations, giving you 36 combinations of style and color to choose from (#238). (#987)

• The tab_stub_indent() function was added to allow for fine control over row label indentation. (#999, #1002)

• The fmt_duration() function is now available for formatting values to time durations whether those input values are numbers or of the difftime class. It provides a wide array of options so you can get the formatting of time duration values just right (#403). (#916)

• Generating PNG files from gt tables using gtsave() is now way better due to the move away from webshot to its sequel package webshot2 (#621, #885). (#985)

• A number of accessibility enhancements for HTML table outputs were undertaken to make gt one of the most accessible table packages around (#678, #679, #680). (#977, #978, #979, #981; thank you, @jooyoungseo, for all of this)

• Errors and warnings, though we don't often like to receive them, look much better now since we've integrated the cli package into gt! (#963)

Minor improvements and bug fixes

• We can now supply data from adjacent columns to user-supplied aggregation functions in summary_rows() and grand_summary_rows() (#383, #690, #952). (#1018)

• Initial alignments in the table stub are now more intelligently determined in gt() by examining the content of the column (the "rowname_col") that will comprise the stub. If values are predominantly number-like then a right alignment will be chosen, otherwise row labels will be left aligned. (#999)

• Enhanced automatic column alignment (set to "auto" by default in the gt() function) by probing the contents of the columns. This ensures that string-based dates/times/datetimes receive a better appearance before formatting. (#997)

• gt tables no longer break table numbering in R Markdown and Quarto documents. (#1003, thanks @cscheid!)

• The as_latex() documentation was updated to provide guidance on necessary LaTeX dependencies. (#1022, thanks @coatless!)

• Fixed a bug where footnote marks would not be rendered in the stub column representing row groups (#1001). (#1014)

• Using gtsave() in a Quarto or R Markdown code chunk will no longer generate any output, which is good since we don't want any (#974). (#1011)

• Fixed an issue where footnote marks would not be set directly to cell text generated from fmt_markdown() (#893, #993). (#1013)

• All testthat tests have been migrated to the 3rd Edition of testthat. Also, the tests run much faster now. (#959)

v0.6.0

New features

• RTF output support is now more suitable for Pharma users: (1) table styling now defaults to a look with far fewer table borders, which is more common for Pharma-based applications; (2) RTF table generation can make use of the new page-layout options (page.*) added to the tab_options() function (e.g., page.orientation, page.numbering, etc.) (#311, #653). (#927)

• The tab_footnote() function now allows for non-location footnotes by making the locations argument optional; any footnotes declared in this way will have their notes precede any location-based notes in the footer area (#879). (#925)

• The fmt_partsper() function was added so that we can format values as high-ratio, dimensionless quantities (e.g., "per-mille", "ppm", etc.). These are less common that percentage values but no less important (#608). (#914)

• The sub_*() group of functions was added to make simple and common substitutions of data in a gt table with not-too-much effort. The functions added are: sub_missing(), sub_zero(), sub_small_vals(), and sub_large_vals(). The sub_missing() function replaces fmt_missing() (and this deprecates the latter) (#904, #910, #911). (#912)

v0.5.0

New features

• The tab_spanner() and tab_spanner_delim() functions now support the addition of column spanners with unlimited levels (previously, only a single level of spanner column labels was possible to add) (#746). (#905)

• There is now support for the Indian numbering system, making the placement of digit-grouping separators correspond to thousand, lakh, crore, and higher quantities. This is available for several fmt_*() functions (fmt_number(), fmt_integer(), fmt_currency(), fmt_percent(), and fmt_fraction()) and is exposed as the system argument with two options "intl" (international, the default) and "ind" (Indian).

Minor improvements and bug fixes

• Enhanced the cols_merge_uncert() function such that different lower and upper uncertainty values could be used (by specifying two columns in the col_uncert argument) (#884). (#888)

• Honor the locale setting from the initial gt() call when rendering summary rows (#906). (#907)

• Fixed an issue where styles that spanned multiple columns and rows in a summary block were not fully rendered (#859). (#924)

• Improved LaTeX rendering of certain characters (e.g, \, ~, &, etc.). (#903)

• Aggregations with summary_rows() and grand_summary_rows() that might return an NaN (common case is using mean(., na.rm = TRUE) on an all-NA column) now return NA. This removes the evaluation error that would otherwise occur and provides an opportunity to replace the NA with an explanatory replacement string (#471, #827). (#887)

v0.4.0

New features

• We now have the ability to display row group labels as a column in the stub (as an alternative to these grouping labels appearing as header rows). This is made possible with the new option row_group.as_column in tab_options() (the default keeps the preexisting behavior) (#291). (#855)

• The table footer (which can contain footnotes and source notes) now has numerous layout options in tab_options(): footnotes.multiline, footnotes.sep, source_notes.multiline, and source_notes.sep. With the *.multline options, we can control whether notes span across separate lines (TRUE, the default) or are kept together as a paragraph (FALSE). For the latter layout option, we can control the separation between consecutive notes with the *.sep options (the default is " "). The work done here also fixes long-standing issues with footer output in the LaTeX and RTF formats (#593, #833). (#864)

• The fmt_fraction() formatter was added, allowing for flexible formatting of numerical values to mixed fractions of configurable accuracy (#402). (#753)

• Added the opt_horizontal_padding() and opt_vertical_padding() functions to easily expand or contract an HTML table in the horizontal and vertical directions (#868). (#882)

• There is now a locale argument in the gt() function. If set, formatter functions like fmt_number() will automatically use this global locale while formatting. There also remains the the option to override the global locale with any non-NULL value set for locale within a fmt_*() call (#682). (#866)

Minor improvements and bug fixes

• There is now more flexibility, improved documentation, and more testing/reliability for the date/time formatting functions (fmt_date(), fmt_time(), and fmt_datetime()). Now, Date and POSIXct columns are allowed to be formatted with these functions. With fmt_datetime(), we can even supply a format code for generation of custom dates/times (#612, #775, #800). (#801)

• Footnote marks for HTML tables now have an improved appearance. They are slightly larger, set better against the text they follow, and, asterisks are specially handled such that their sizing is consistent with other marks (#511). (#876)

• Further improving support for color value inputs, gt now allows shorthand hexadecimal color values (like #333) and the use of the transparent CSS color keyword (#839, #856). (#870)

• The rendering of percent signs in HTML output was changed from the problematic HTML entity % to simply %. This solves issues with tables incorrectly rendering percent signs in some situations (#657, #686). (#871)

• All values supplied within cells_title() are now checked more rigorously, erroring and providing actionable messaging when stopping does occur (#424). (#783)

• Added a missing column entry in the documentation for the gtcars dataset. Thanks, @luisDVA for providing this fix! (#883)