single point `scatter3d` traces fail in R

[cldougl]

library(plotly)
p<-plot_ly(
  x = 1,
  y = 1,
  z = 1,
  type = "scatter3d", 
  mode = "markers")
p

plotly_POST(p)

[etpinard]

@cldougl really?

That shouldn't work. x, y, and z need to be arrays unless plotly_POST does something fancy.

[cpsievert]

@etpinard yea, I forgot about "z". This will be an easy fix.

> plotly:::get_boxed
function() {
  c("x", "y", "lat", "lon", "text")
}

[cpsievert]

@etpinard while were on the topic can you think of any other properties that should always be boxed, even if length 1?

[etpinard]

@cpsievert I guess all 'data_array' attributes as listed in the schema

e.g : https://github.com/plotly/plotly.js/blob/master/src/traces/scatter3d/attributes.js#L63

[cpsievert]

Thanks! Just for posterity:

$ cd plotly.js/src/traces
$ ag -B 1 'data_array'
box/attributes.js
20-    y: {
21:        valType: 'data_array',
27-    x: {
28:        valType: 'data_array',

choropleth/attributes.js
19-    locations: {
20:        valType: 'data_array',
27-    z: {
28:        valType: 'data_array',
31-    text: {
32:        valType: 'data_array',

choropleth/plot.js
43-
44:        // 'data_array' attributes

heatmap/attributes.js
18-    z: {
19:        valType: 'data_array',
28-    text: {
29:        valType: 'data_array',

histogram/attributes.js
20-    x: {
21:        valType: 'data_array',
26-    y: {
27:        valType: 'data_array',

histogram2d/attributes.js
18-    z: {
19:        valType: 'data_array',
23-        color: {
24:            valType: 'data_array',

mesh3d/attributes.js
17-    x: {
18:        valType: 'data_array',
21-    y: {
22:        valType: 'data_array',
25-    z: {
26:        valType: 'data_array',
30-    i: {
31:        valType: 'data_array',
34-    j: {
35:        valType: 'data_array',
38-    k: {
39:        valType: 'data_array',
68-    intensity: {
69:        valType: 'data_array',
82-    vertexcolor: {
83:        valType: 'data_array',  // FIXME: this should be a color array
90-    facecolor: {
91:        valType: 'data_array',

pie/attributes.js
19-    labels: {
20:        valType: 'data_array',
42-    values: {
43:        valType: 'data_array',
48-        colors: {
49:            valType: 'data_array',  // TODO 'color_array' ?
80-    text: {
81:        valType: 'data_array',

pie/layout_attributes.js
16-     */
17:    hiddenlabels: {valType: 'data_array'}

scatter/attributes.js
17-    x: {
18:        valType: 'data_array',
41-    y: {
42:        valType: 'data_array',
424-    r: {
425:        valType: 'data_array',
431-    t: {
432:        valType: 'data_array',

scatter3d/attributes.js
54-    x: {
55:        valType: 'data_array',
58-    y: {
59:        valType: 'data_array',
62-    z: {
63:        valType: 'data_array',

scattergeo/attributes.js
21-    lon: {
22:        valType: 'data_array',
25-    lat: {
26:        valType: 'data_array',
29-    locations: {
30:        valType: 'data_array',

surface/attributes.js
82-    z: {
83:        valType: 'data_array',
86-    x: {
87:        valType: 'data_array',
90-    y: {
91:        valType: 'data_array',
94-    text: {
95:        valType: 'data_array',

[federicou]

actually I'm encountering the same issue:

data.frame(c=c(rep("A",32),rep("B",2),rep("C",1)),x=runif(35),y=runif(35),z=runif(35)) %>%
plot_ly(x=x,y=y,z=z,color = c, type="scatter3d", mode="markers") -> p3

The plot displays only the 2 categories that have at least 2 points