Description
Box, violin, candlestick and ohlc traces ask their position axis for a minimum tick spacing, so that each box gets a tick of its own rather than ticks at meaningless in-between positions. That forcing is silently lost whenever the figure is drawn into a graph div that already held a different figure — so Plotly.react, restyle, addTraces and friends can produce different ticks than Plotly.newPlot of the exact same figure.
It goes the other way too: a graph div that once held a box plot keeps the forced spacing after you react to a trace type that never asked for it.
Steps to reproduce
var boxFig = {
data: [{type: 'box', x: [1, 1, 2, 2, 3, 3], y: [1, 2, 3, 4, 5, 6]}],
layout: {width: 700, height: 400}
};
var scatterFig = {data: [{y: [1, 2, 3]}], layout: {width: 700, height: 400}};
// x ticks: 1, 2, 3 — correct, one per box
Plotly.newPlot(gd, boxFig);
// x ticks: 0.5, 1, 1.5, 2, 2.5, 3, 3.5 — same figure, half the ticks land between boxes
Plotly.newPlot(gd, scatterFig).then(function() { return Plotly.react(gd, boxFig); });It is most visible on a date axis. Five daily candles at width: 1000, reached by reacting from a line chart over the same dates:
newPlot: Jan 1 | Jan 2 | Jan 3 | Jan 4 | Jan 5
react: 12:00 Dec 31, 2023 | 00:00 Jan 1, 2024 | 12:00 | 00:00 Jan 2, 2024 | 12:00 | ...
Every second tick falls in the gap between two candles, and the axis now starts on the day before the data.
Notes
Axes.minDtick keeps its state in ax._minDtick / ax._forceTick0, with three meanings: undefined (nothing forced yet — adopt this trace's spacing), a positive number (a forcing is in effect), and 0 (forcing cancelled — by non-grouped bars, or by scatter/heatmap calc — and sticky, so a later trace can't reinstate it).
The reset back to undefined between passes sits at the bottom of setConvert, but it can't take effect on a real axis: setConvert runs while supplyDefaults builds the new _fullLayout, where those keys don't exist yet, and relinkPrivateKeys then copies the old values back onto it. So whichever figure the graph div held first decides the forcing for every figure after it. newPlot escapes it only because it starts from an empty _fullLayout.
Found by a sweep comparing react against newPlot of the same figure — 7 of 12 divergences traced back here.
Fix proposed in #7950, which moves the reset into ax.clearCalc(), alongside the equivalent relink cleanup doCalcdata already does for shared color axes.
Description
Box, violin, candlestick and ohlc traces ask their position axis for a minimum tick spacing, so that each box gets a tick of its own rather than ticks at meaningless in-between positions. That forcing is silently lost whenever the figure is drawn into a graph div that already held a different figure — so
Plotly.react,restyle,addTracesand friends can produce different ticks thanPlotly.newPlotof the exact same figure.It goes the other way too: a graph div that once held a box plot keeps the forced spacing after you react to a trace type that never asked for it.
Steps to reproduce
It is most visible on a date axis. Five daily candles at
width: 1000, reached by reacting from a line chart over the same dates:Every second tick falls in the gap between two candles, and the axis now starts on the day before the data.
Notes
Axes.minDtickkeeps its state inax._minDtick/ax._forceTick0, with three meanings:undefined(nothing forced yet — adopt this trace's spacing), a positive number (a forcing is in effect), and0(forcing cancelled — by non-grouped bars, or by scatter/heatmapcalc— and sticky, so a later trace can't reinstate it).The reset back to
undefinedbetween passes sits at the bottom ofsetConvert, but it can't take effect on a real axis:setConvertruns whilesupplyDefaultsbuilds the new_fullLayout, where those keys don't exist yet, andrelinkPrivateKeysthen copies the old values back onto it. So whichever figure the graph div held first decides the forcing for every figure after it.newPlotescapes it only because it starts from an empty_fullLayout.Found by a sweep comparing
reactagainstnewPlotof the same figure — 7 of 12 divergences traced back here.Fix proposed in #7950, which moves the reset into
ax.clearCalc(), alongside the equivalent relink cleanupdoCalcdataalready does for shared color axes.