gtsummary Files

New Features and Functions

• Added tbl_split_by_rows() and tbl_split_by_columns() to split tables horizontally (row-wise) and vertically (column-wise). (#2216)

• Users are now allows to specify/override the denominator by passing an integer or a data frame to the tbl_summary(percent) argument. (#2239)

• Added the tbl_merge(tbl_ids) and tbl_stack(tbl_ids) arguments that allows used to label the gtsummary input tables. This is particularly helpful when calling gather_ard(), which will return a named list of ARDs where the names are the assigned tbl IDs. (#2224)

• Adding new function add_difference_row(). The function is similar to add_difference(), except that differences are placed on the rows below the summary statistics. (#2138)

• The style_*() and label_style_*() functions have gained the na argument allowing users to specify how NA values are returned.

• The tbl_stack(attr_order) argument has been added that allows users to specify the order in which the individual attributes take precedent when they are stacked. For example, to use the headers from the second table, specify attr_order=2. (#2202)

• The tbl_regression() function now takes advantage of a new feature in {broom.helpers} to support multicomponent models, such as, multinomial models and mixed-effects models. Review broom.helpers::tidy_group_by() for details. (#1540)

Other Updates

• Added an article on modify_*() functions. (#2209)

• Added theme elements tbl_hierarchical_count-fn:addnl-fn-to-run and tbl_hierarchical_count-fn:addnl-fn-to-run. (#2262)

• Added the "add_overall.tbl_summary-arg:col_label" theme element to control the default value of add_overall.tbl_summary(col_label). (#2246)

• Variable labels are now retained with tbl_summary(sort=~'frequency'). (#2260)

• Functions sort_hierarchical() and filter_hierarchical() are now S3 generics.

• The add_overall() function no longer returns an error when an unstratified table is passed. Rather, a message is printed and the unaltered table is returned.

Lifecycle Updates

• Updated the function name of modify_column_indent() to the more accurately named modify_indent() as the function operates on cells rather than columns.

Bug Fixes

• Fix in tbl_svysummary() when there are missing values in the by column.

• Fixed bug in tbl_hierarchical() causing the overall row label to be incorrectly used as a variable label.

• Fixed bug in sort_hierarchical() causing some rows to be dropped when sorting unstratified tables. (#2237)

• Fixed bug in filter_hierarchical() causing an error for unstratified tables.

• Fix in gather_ard() for tbl_regression() and tbl_uvregression(). Previously, only the ARD for the primary regression model(s) would be returned, and now all ARDs are returned including those from subsequent calls to add_*() functions. (#2208)

• Fix in tbl_merge(merge_vars) argument when a table contained a "row_type" column and the tables were not merged by this variable (which is included in the default). (#2205)

• Fix in add_overall() where not all table styling were copied from the overall table to the newly merged table. This did not affect any native gtsummary tables, but did appear in an extension package.

• Fix in experimental function modify_post_fmt_fun() when tables including this call, did not specify the rows argument, and were later stacked. The enquo environment associated with the rows argument was empty when the argument is not specified, and the empty environment caused an issue with evaluation if the table was later stacked.

• Fix in tbl_svysummary(missing='ifany') when the weighted number of missing observations was <= 0.5. Previously, these counts were coerced to integer and rounded to zero, and therefore, did not appear in the table. (#2229)

• Fix in gather_ard() for tbl_strata_nested_stack() tables, where the function returned an empty list. (#2223)