Why StyledTables.jl?

DOCX is a common output format in clinical and applied research, and existing packages either did not support docx or did not provide a high level interface for custom table styling.

Existing table-styling packages

I tried to make the tables I need with PrettyTables.jl and SummaryTables.jl, but neither met my requirements.

The building blocks SummaryTables provides work well for DOCX rendering, but the tables I needed were often outside the intentionally opinionated summaries this package offers.

PrettyTables, on the other hand, offered the customisation I needed, but had no DOCX support.

What was missing

For me the missing piece was a simple, intuitive way to combine the building blocks provided by SummaryTables. R's gt package is the clearest model for how this works: construct a table object, then apply modifier functions one at a time.

StyledTables follows the same incremental approach: each step is a small, focused modification, and you can inspect the result at any stage.

Unlike R's idiomatic pipe syntax, StyledTables has a more Julian take on iterative table modification: all styling functions are mutating and modify the StyledTable in place.

The workflow

In contrast to SummaryTables, StyledTables provides no data-summarising capabilities. It assumes you bring summarised data that can be converted to a DataFrame. You typically go through the following steps when working with StyledTables:

1. Summarise your data

2. Convert it to a StyledTable

3. Apply styling

4. Render to HTML, LaTeX, Typst, or DOCX