NEWS.md
To support creating identical PDF output on rerun, st2report(), st2article(), and st2pdf() now omit some PDF metadata, and st2report() no longer includes the date in its footer (#402).
By default, st2article() and st2doc() no longer try to display the generated PDF in a non-interactive context (#397).
st2article() and st2report() should hide the underlying build output by default, but, with the switch to latexmk in the last release, output made it through in some cases. Standard error from the underlying latexmk process is now discarded by default to restore the intended default behavior (#399).st2article(), st2report(), st_aspdf(), and st_aspng() is now built with latexmk, which reruns the tex engine as many times as needed; these functions now require latexmk to be installed and in your PATH. The ntex argument is deprecated and ignored; passing it results in a warning (#393).stable_long() where the \hline at the bottom of the table was misplaced; tables now get a closing \hline on the final page for both single- and multi-page outputs (#391).stable_long() where extraneous \hline commands were written at the start and end of the table body and a stray \endhead was written immediately after \begin{longtable} (#388, #389).sig() no longer renders numbers in scientific notation by default; pass maxex to opt into scientific notation for numbers with exponents at or above that threshold (#378).
sig() gains a big.mark argument (following formatC) to insert a thousands separator; the mark is applied only when the formatted number has 5 or more digits (#378).
pt_cont_wide() and pt_cont_long() now ungroup incoming data before summarizing, preventing spurious “Adding missing grouping variables” messages (#375).
The ID column used by data inventory and related functions can now be overridden globally via the mrg.id_col option (#380).
The deprecated id_col argument to pt_cont_wide() and pt_cont_long() has been removed (#380).
Global options are now documented and accessible via ?pmtables (#381).
pt_data_inventory() where missing values in the by column resulted in the missing value summary row getting labeled as an “All data” summary row; this fix replaces missing values with the character string "NA" as the table is built and “All data” summary rows are now properly labeled (#373).\lbrack or \rbrack; this behavior can be modified or turned off through an argument called mask_bracket which is passed to make_tabular() (masking brackets in the table body) or tab_notes() (masking brackets in table notes); both make_tabular() and tab_notes() can receive the mask_bracket argument through stable() or stable_long() (#367).shortplural or longplural via read_glossary() ; parsing is now done through tools::parseLatex (#370).pt_demographics() was refactored (#363).
Statistic is “Missing” (provided by the default summary function for continuous data) and all summary values are “0” are dropped by defaultdrop_miss can be used to optionally retain the “Missing” stat for all continuous data summariespt_inventory_long() is a new function to create data inventory summaries like those generated by pt_data_inventory(), but stratified by multiple covariates (#360).
When a caption macro name is provided to stable_long(), it will only be used if the macro is defined; if the macro is not defined, some replacement text is included noting that the macro will be used once it is defined (#355).
st_as_image() and st2report()) where a list was getting passed as .envir to glue::glue() rather than an environment; this was causing glue::glue() to fail after changes rolled out in glue version 1.8.0 (#346).read_glossary() reads from .tex or .yaml formatted files, returning a glossary objectselect_glossary() selects specific entries from a glossary objectas_glossary() creates a glossary object on the fly or from a listglossary_notes() creates notes from a glossary object, a list, or the name of a glossary filest_notes_glo() creates and and attaches notes to a table in a pipelinestable_save_image() added to create an image from a table and save to a specific location; this builds on st_aspdf() and st_aspng(), adding convenient syntax and options (#333).maxex argument to sig() can now be set through the pmtables.maxex option (#328).pt_cat_long() when the by argument is used with no all data summary (#330).st_caption() function added to associate a caption (including short caption and possibly a label) to any table; captions will appear in st2report() outputs; captions will always be written to file for stable_long outputs and can be optionally written to file for stable outputs (#313).
yaml_as_df() will now retain row names in a column named by the new argument row_var (#317).
st_align(), st_left(), st_right() and st_center() can now be called multiple times in a pipeline, supported by new update method for aligncol objects (#315).
st_clone() function added to make a copy of stobject objects (#314).
The id_col argument to pt_cont_long() and pt_cont_wide() is deprecated (#305).
sumrows was not recognized as a valid field in stobject objects (#312).New function st_filter() to filter data item in a pipeline (#298).
Add summarize_all and all_name_stacked arguments to pt_data_inventory() (#297).
Fixed bug where all_name was not getting used in pt_data_inventory() (#297).
Fixed bug where detached table notes were getting rendered too close to the main table when building standalone pdf under certain TeX distributions (#286).
New functions st_as_image(), st2pdf(), and st2png() to render tables with TeX to either pdf or png format; image files may be kept on disk or read back for display while knitting; pdflatex system dependency for pdf images and latex + dvipng for png images; additional Suggested packages include magick and pdftools (#277, #278).
sig() now returns character when integer type is passed (#272).
st_new() is now generic with dispatch for data.frame and objects with class pmtable; most pipeline functions can now be used to customize tables coming from pt_cat_*, pt_cont_*, pt_demographics() and pt_data_inventory() (#274).
New functions st_notes_detach(), st_notes_rm(), st_notes_app(), st_notes_str(), st_notes_sub() and st_notes_conf() to help working with table notes in a pipe context (#274).
st_span() and st_span_split() accept align argument to push column spanner titles to the left, right or center (default); the argument eventually gets passed to colgroup() (#261).
pt_cat_wide(), pt_cat_long() and pt_demographics() gain argument denom to alter the denominator when calculating percents for categorical data summaries (#268).
pt_cat_wide() gains argument complete to display missing levels of by and panel (#268).
Put stories in yaml format; add script to build validation docs from the yaml file (#269, #270).
colgroup() (and st_span()) gains an align argument to position the spanner title on left or right in addition to the center (default) #260, #261.
rowpanel() (and st_panel()) gains jutargument to push non-panel table contents to the right relative to the panel header row so that contents under the panel header are indented #251, #253.
Panel header rows are now modified so that the header row stays with the first non-header row for longtable output #252, #253.
Consistent BQL / BLQ handling for column titles and table notes for pt_data_inventory() #254, #255.
Add cols_omit option to omit column header data (#213)
Add pt_demographics function to generate a new table of both continuous and categorical data (#186, #249)
Add title_side argument to colsplit() so that the title can be taken from left or right side of split (#231)
Add hline argument to rowgroup() constructor to make the horizontal line above the panel data optional (#215)
Refactor pt_data_inventory() to calculate percent BLQ using denominator that is the sum of the number of observations BLQ and non-BLQ / non-missing (#221, #222)
List names are now escaped when passing a list of tables to st2report() and friends (#232)
Add newline after printing table text using st_asis() (#224)
Fix bug where arguments could not be passed along to stable_long() when coercing pmtable output (#203)
Fix bug where user-specified font size was not getting propagated to long table output with stable_long() (#204)
.list argument to st_rename() allowing user to pass rename information as a named list with format old name = new name #189drop_miss wasn’t executing correctly #182stable_long() is now a generic function with methods for data.frame, stobject and pmtable #179st_span() was slightly refactored so that it would dispatch to st_span_split() when the split argument was passed #173span_split() was updating column names, but this was happening before other spans were calculated #172NA) and will replace them with the string "NA" and issue a warning #117st2report(), st2article() or st2doc(); see landscape argument to st_wrap() #129lt_cap_short argument to stable_long() to set a short table title that will appear in list of tables #131na_fill argument to set placeholder #133hline indices when set with st_hline() #134cols_bold that prevented sanitizing text prior to styling as bold #141st2report() to quickly preview tables in a real-like report document, build directly with pdflatex #148st2article() to quickly preview tables in a latex article-like, built directly with pdflatex #148st_asis() function and asis argument to st_wrap(); pipe tables to st_asis() or use asis argument to render tables in line in an Rmd document; in addition to formatting output “as-is”, latex dependencies are also invoked #149yaml_as_df() functionality to accept table columns by position, rather than name #154stable() method for stobject #159by argument to pt_cont_long so that, when panel is also passed, the table is paneled by col and is “by” the panel variable #92pt_cat_long so that summary numbers are shown underneath column labels #102row and col in tab_size() with the same behavior as tabular / stable() #105cols_extra feature, allowing extra column header information passed as a data frame (see tab_cols()) #102, #118data arguments to reinforce that pmtables doesn’t filter or subset data prior to processing or summarizing #120stable_long() where stable_file wasn’t getting savedn summary column to pt_cat_wide() #80pt_wrap() generic; add method for stable_long that doesn’t wrap with table #121duplicates_ok argument to rowpanelduplicates_ok is FALSEget_debug_data()
st_* functions that can be used in a pipeline to set arguments for stable(); see inst/demo-pipe.pdf
st2doc will render multiple page document when a list of tables is passed in as textstable argument from bold_cols to col_bold
col_bold argumentr_file and output_file are only settable with options r.file.label and output.file.label, respectivelycol_space, row_space and fontsize arguments to stable(); they should be set with a call to tab_size
st_sizes() to set sizes argument to stable()
stable() function in order to reduce cyclomatic complexity and shorten overall function length in response to brief review and comment by TS representativepanel argument as a rowgroup objectas.panel() function for constructing rowgroup objectspanel is passed with an alternate name, the name is is used to set the prefixpmtables.dir to set output directory for stable_save()
mrg.script to set script name in st_files() and tab_notes()
... to create multi-line column namesheader_row argument to tab_cols() to control the spacing between column labels and between labels and units?pt_opts
center argument to col_fixed; when ragged is no, the column can still be fixed and centered...) added to st_new() / st_data() to be collected by argsst_units() pipeline functiontex_bold() function to style data frame contentstex_it() function to style data frame contentsst_edit()
st_hline(); allow to add hline by regexp across whole data frame or targetedst_bold and st_it to target columns or whole data framest_drop()
analysis1, pmt_pk, pmt_first, pmt_obs
tab_cols: all col_xyz arguments are renamed to cols_xyz
find_bq_col function to allow either BQL or BLQ as default value for bq_col argument to pt_data_inventory
pattern and cols arguments to st_sumrow; this allows calculation of rows based on a regular expressiondf_grep_rows and df_grepl_rows functions for finding rows where a regular expression is matchedpmtables.escape option to pass a character vector of items to be escapedheader_row (tab_size()) is a negative number; enter a negative number if you want to remove space between header rows; this should only be required when row space is increased