Interrupt (null values) as gap mode for stacked scatter charts

[Bjodol]

Hi!

We have been evaluating a lot of charting frameworks and found that plotly is the one that fit our needs the best. However we have one major issue with it on the stacked scatter charts. Currently the only two supported gapModes is "infer zero" and "interpolate" which hinders us from fully commit to Plotly. In our application a null value signifies a data loss and replacing it with 0 provides a wrong picture as the 0 indicates that we have data and the value is 0 not that we actually in fact have a data loss.

I pulled down the code and saw that src/traces/scatter/plot.js:407 suggests interrupt as a future new mode. Is this covering null values not being shown in the area charts and is there an ETA for when this will arrive? If interrupt is something else, I would like to propose having a gap mode where null values aren't drawn in the stacked group.

[alexcjohnson]

Thanks @Bjodol, there definitely should be an issue open for this. That's right, we had discussed from the beginning that a third option to stackgaps making it clearer that a gap is really "missing" would be useful. There's also this commented-out portion in the attribute description:

plotly.js/src/traces/scatter/attributes.js

Lines 138 to 143 in 2ceb5c3

	'With *infer zero* we insert a zero at these locations.',
	'With *interpolate* we linearly interpolate between existing',
	'values, and extrapolate a constant beyond the existing values.'
	// TODO - implement interrupt mode
	// '*interrupt* omits this trace from the stack at this location by',
	// 'dropping abruptly, midway between the existing and missing locations.'

which, as mentioned when we were originally discussing this feature would look something like:

Is that the kind of behavior you had in mind?

We don't have an ETA for this currently, but it's useful to know folks are interested. But the best way to get it to happen quickly would be either to submit a PR or to sponsor the project https://plot.ly/products/consulting-and-oem/

[Bjodol]

This is exactly the behavior I had in mind yes! I hope my interest would spark some priority on it. But knowing it is planned will probably make us go for plotly. Would love to submit a PR to it myself, but currently with hanging deadlines I cannot really invest the time in it now. And honestly I wouldn't even know where to start 😂

[warren-oneill]

Hi @Bjodol and @alexcjohnson, I just sent in a request to sponsor this issue getting done here https://plotly.com/get-pricing/.
I have a similar issue with intermittent timeseries data. As you can see in the plot below I'm getting a lot of zeros in the hover that I don't want to see.

[warren-oneill]

However, I was able to remove all the zeroes from the plot by setting mode='none'. It's only the issues with the hover I am having.

[warren-oneill]

I ended up getting around the issues by using a stacked bar plot with no gaps or borders.

fig = go.Figure()
fig.add_trace(go.Bar(x=x, y=y, name=u, opacity=0.5, 
                                    marker=dict(line=dict(width=0))))
fig.update_layout(barmode='stack', bargap=0)

[gvwilson]

Hi - this issue has been sitting for a while, so as part of our effort to tidy up our public repositories I'm going to close it. If it's still a concern, we'd be grateful if you could open a new issue (with a short reproducible example if appropriate) so that we can add it to our stack. Cheers - @gvwilson