Summary
Currently, when no hatch argument is passed to Axes.grouped_bar(), hatching is entirely disabled:
if hatch is None:
# No hatch specified: disable hatching entirely by cycling [None].
hatches = itertools.cycle([None])
# TODO: Discussion —
# Should grouped_bar() apply a default hatch pattern (e.g., '//')
# when none is provided ?
This behavior matches the philosophy of “explicit is better than implicit,” keeping hatching opt-in.
However, there’s a possible usability question:
Should grouped_bar() automatically apply a default hatch pattern (for example '//') when multiple datasets are plotted, to help users visually distinguish bars even when colors are similar or when figures are printed in grayscale?
Discussion Points:
Would automatic hatching improve accessibility and readability (especially in black-and-white printing)?
Should the default depend on the number of datasets or color distinctness?
Follow-up to the Enh/Add hatch pattern support to Axes.grouped_bar #30726
Proposed fix
No response
Summary
Currently, when no hatch argument is passed to Axes.grouped_bar(), hatching is entirely disabled:
This behavior matches the philosophy of “explicit is better than implicit,” keeping hatching opt-in.
However, there’s a possible usability question:
Should grouped_bar() automatically apply a default hatch pattern (for example '//') when multiple datasets are plotted, to help users visually distinguish bars even when colors are similar or when figures are printed in grayscale?
Discussion Points:
Would automatic hatching improve accessibility and readability (especially in black-and-white printing)?
Should the default depend on the number of datasets or color distinctness?
Follow-up to the Enh/Add hatch pattern support to Axes.grouped_bar #30726
Proposed fix
No response