Swap the rows and columns of a cross-table

Turns a table's rows into its columns and its columns into its rows, and returns a real tabxplor_tab — one you can keep piping through dplyr, colour, print and export. Row percentages become column percentages, and the old total column and total row swap places.

Its job is the profile table: many variables down the page, a few groups across it. It is also the only way to put a mean on a row, since a number given to row_vars is always cut into levels — build the means as columns (tab(data, groups, numeric_vars)) and transpose.

Use the exporters' transpose = TRUE argument instead whenever you only need the OUTPUT: it flips the finished render model after every colour and cell string is computed, so it handles what a data-level flip cannot (several row variables, tab_vars sub-tables, columns of unlike kinds).

tab(data, row_vars, col_vars, pct = "row") |> tab_kable(transpose = TRUE)   # or tab_md() / tab_xl()

Usage

tab_transpose(tabs, name = NULL)

Arguments

tabs

A single table made with tab (one row variable, one column variable; not a subtabled table with tab_vars, and at most one total row and one total column).

name

The name to give the new first (label) column, holding the old column-variable levels. NULL (default) uses the old column-variable name.

Value

A transposed tabxplor_tab.

Columns of unlike kinds

A transposed column stacks whatever the original rows held, so transposing a table that mixes percentage and mean columns gives a mixed column. Its numbers and its cell layouts are exact; only the colour ladder is shared, so an additive measure (color = "difference") grades the percentage cells and leaves the others uncoloured, while a multiplicative one (color = "ratio") grades them all — the percentage and mean ratio ladders being the same rungs.

See also

tab_spread, tab_compact, tab_export.

Examples

# race x marital, read as marital x race:
tab(forcats::gss_cat, marital, race, pct = "row") |> tab_transpose()
#> # A tabxplor tab: 5 × 8
#>   race   `No answer` `Never married` Separated Divorced Widowed Married  Total
#>               <col%>          <col%>    <col%>   <col%>  <col%>  <col%> <col%>
#> 1 Other          12%             12%       15%       6%      4%      9%     9%
#> 2 Black          12%             24%       26%      15%     14%      9%    15%
#> 3 White          76%             64%       59%      79%     82%     82%    76%
#> 4 Total         100%            100%      100%     100%    100%    100%   100%
#> 5 n               17           5 416       743    3 383   1 807  10 117 21 483

# \donttest{
# the profile table: mean rows come from mean columns
tab(forcats::gss_cat, marital, c(age, tvhours)) |> tab_transpose()
#> # A tabxplor tab: 2 × 8
#>   age      `No answer` `Never married`     Separated     Divorced      Widowed
#>            <mean (cv)>     <mean (cv)>   <mean (cv)>  <mean (cv)>  <mean (cv)>
#> 1 age      52 (cv 32%)     34 (cv 40%)  45 (cv  30%)  51 (cv 26%)  72 (cv 18%)
#> 2 tvhours 2.6 (cv 44%)    3.1 (cv 92%) 3.5 (cv 101%) 3.1 (cv 89%) 3.9 (cv 74%)
#> # ℹ 2 more variables: Married <mean (cv)>, Total <mean (cv)>
# }