Plotly.js: "bargroupgap" adds space between histogram "bars"

Created on 4 Apr 2018  路  10Comments  路  Source: plotly/plotly.js

Most helpful comment

I vote for _not a bug_.

I think so too, I'm afraid. You can make a histogram on a categorical axis, for example, which is of course not continuous and therefore a gap is reasonable. And you can mix (and stack) histogram traces with bar traces for cases where some of your data is pre-binned, and in that case it doesn't make sense to have two different gaps, we would have to pick one of them to win and ignore the other.

All 10 comments

Right. Histograms have bargroupgap set to 0 by default. So, you're arguing that bargroupgap should be ignored completely always for histogram traces?

Exactly, because it's continuous instead of categorical. What do you think?

because it's continuous instead of categorical

Yep, that's why be default bargroupgap: 0. But if the user really _really_ wants gaps in there, I'm not sure we should stop them.

At the very least what you're asking is a breaking change, so it won't happen anytime soon.

Understood :) I'm making a theme for Plotly and try to generalize as much as possible. The easiest would be to use one master layout for all charts.

I'd add another property like histogramgab. That wouldn't be a breaking change.

I'd add another property like histogramgab. That wouldn't be a breaking change.

That's an idea, but still it would be a breaking change. Some users might be using bargroupgap on histograms right now.

That's correct. Just for me to know for my personal future work: Is it considered a bug by you or will it never be changed?

I vote for _not a bug_.

By the way, if you're interested in _themes_, you may want to follow https://github.com/plotly/plotly.js/issues/2469

I vote for _not a bug_.

I think so too, I'm afraid. You can make a histogram on a categorical axis, for example, which is of course not continuous and therefore a gap is reasonable. And you can mix (and stack) histogram traces with bar traces for cases where some of your data is pre-binned, and in that case it doesn't make sense to have two different gaps, we would have to pick one of them to win and ignore the other.

Ah great and very relevant. Closing this.

Was this page helpful?
0 / 5 - 0 ratings