Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Score big with last-minute savings on the final tickets to FabCon Vienna. Secure your discount

Reply
jonespossibly
Advocate I
Advocate I

Another Deneb query - 2 data fields on one bar chart

I have cause to render two different data fields on to the same bars in the same chart.

so for instance the x axis will be date

y axis should have field 'A' and field 'B' stacked on top of each other for the given date.

Also both fields are grouped by another field, 'Z'.

Groupings for A should be shown separately from Groupings for B in the legend,

 

Getting A to work is simple:

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "values":[
      {"A":10, "B":20, "Z":"poop", "created":"2022/01/01"},
      {"A":10, "B":10, "Z":"moop", "created":"2022/01/01"},
      {"A":10, "B":5, "Z":"doop", "created":"2022/01/01"},
      {"A":13, "B":12, "Z":"poop", "created":"2022/02/01"},
      {"A":33, "B":10, "Z":"moop", "created":"2022/02/01"},
      {"A":0, "B":5, "Z":"doop", "created":"2022/02/01"},
      {"A":13, "B":5, "Z":"poop", "created":"2022/03/01"},
      {"A":3, "B":2, "Z":"moop", "created":"2022/03/01"},
      {"A":20, "B":7, "Z":"doop", "created":"2022/03/01"}
      ]
    },
  "mark": "bar",
  "encoding": {
    "x": {
      "timeUnit": "month",
      "field": "created",
      "type": "ordinal"
    },
    "y": {
      "field":"A",
      "type": "quantitative"
    },
    "color": {
      "field": "Z",
      "type": "nominal"
    }
  }
}

how would i go about adding the figures from 'B' onto each bar, in different colours and presented on the legend on separate rows?

1 ACCEPTED SOLUTION

Here's another approach, using a single legend, but using a fold transform to unpivot, and a calculate transform to assign a unique 'grouping' value that can be coloured accordingly. This solution uses a single mark also:

dmp_1-1650324486236.png

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "values": [
      {"A": 10, "B": 20, "Z": "poop", "created": "2022/01/01"},
      {"A": 10, "B": 10, "Z": "moop", "created": "2022/01/01"},
      {"A": 10, "B": 5, "Z": "doop", "created": "2022/01/01"},
      {"A": 13, "B": 12, "Z": "poop", "created": "2022/02/01"},
      {"A": 33, "B": 10, "Z": "moop", "created": "2022/02/01"},
      {"A": 0, "B": 5, "Z": "doop", "created": "2022/02/01"},
      {"A": 13, "B": 5, "Z": "poop", "created": "2022/03/01"},
      {"A": 3, "B": 2, "Z": "moop", "created": "2022/03/01"},
      {"A": 20, "B": 7, "Z": "doop", "created": "2022/03/01"}
    ]
  },
  "transform": [
    {"timeUnit": "month", "field": "created", "as": "month"},
    {"fold": ["A", "B"], "as": ["series", "value"]},
    {"calculate": "datum.series + ': ' + datum.Z", "as": "grouping"}
  ],
  "mark": "bar",
  "encoding": {
    "x": {"timeUnit": "month", "field": "created", "type": "ordinal"},
    "y": {"field": "value", "type": "quantitative"},
    "color": {"field": "grouping"}
  }
}

Editor link

Regards,

Daniel





Did I answer your question? Mark my post as a solution!

Proud to be a Super User!


On how to ask a technical question, if you really want an answer (courtesy of SQLBI)




View solution in original post

3 REPLIES 3
dm-p
Super User
Super User

EDIT: I think that this solution doesn't stack properly and series are overlaid. I'm having a think about that one.

Hi @jonespossibly,

It's not entirely clear what you're asking for - perhaps a sketch or mockup may help for future questions.

I've taken a guess though, and hopefully what I've attempted might still get you there if it's not quite what you were envisioning.

If you want to use the inbuilt legend functionality, you're best splitting into layers and either using two different encoding channels for the colour, (e.g. color and fill), or specifying an independent scale resolution for each color channel so that the legend treats them separately (I've gone for this option in my spec below).

You will also need to specify a range for each color encoding's scale, as they will resolve the same color values if treated as nominal (due to them being the same across series).

I've also specified a manual title for each legend so that it's apparent as to what's what, as Vega-Lite will just title it with the field you're using if not specified:

dmp_0-1650322558002.png

 

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "values": [
      {"A": 10, "B": 20, "Z": "poop", "created": "2022/01/01"},
      {"A": 10, "B": 10, "Z": "moop", "created": "2022/01/01"},
      {"A": 10, "B": 5, "Z": "doop", "created": "2022/01/01"},
      {"A": 13, "B": 12, "Z": "poop", "created": "2022/02/01"},
      {"A": 33, "B": 10, "Z": "moop", "created": "2022/02/01"},
      {"A": 0, "B": 5, "Z": "doop", "created": "2022/02/01"},
      {"A": 13, "B": 5, "Z": "poop", "created": "2022/03/01"},
      {"A": 3, "B": 2, "Z": "moop", "created": "2022/03/01"},
      {"A": 20, "B": 7, "Z": "doop", "created": "2022/03/01"}
    ]
  },
  "resolve": {"scale": {"color": "independent"}},
  "layer": [
    {
      "mark": "bar",
      "encoding": {
        "x": {"timeUnit": "month", "field": "created", "type": "ordinal"},
        "y": {"field": "A", "type": "quantitative"},
        "color": {"field": "Z", "legend": {"title": "Z (A Series)"}}
      }
    },
    {
      "mark": "bar",
      "encoding": {
        "x": {"timeUnit": "month", "field": "created", "type": "ordinal"},
        "y": {"field": "B", "type": "quantitative"},
        "color": {
          "field": "Z",
          "legend": {"title": "Z (B Series)"},
          "scale": {"range": ["red", "green", "blue"]}
        }
      }
    }
  ]
}

 

Regards,

Daniel

 





Did I answer your question? Mark my post as a solution!

Proud to be a Super User!


On how to ask a technical question, if you really want an answer (courtesy of SQLBI)




Here's another approach, using a single legend, but using a fold transform to unpivot, and a calculate transform to assign a unique 'grouping' value that can be coloured accordingly. This solution uses a single mark also:

dmp_1-1650324486236.png

{
  "$schema": "https://vega.github.io/schema/vega-lite/v5.json",
  "data": {
    "values": [
      {"A": 10, "B": 20, "Z": "poop", "created": "2022/01/01"},
      {"A": 10, "B": 10, "Z": "moop", "created": "2022/01/01"},
      {"A": 10, "B": 5, "Z": "doop", "created": "2022/01/01"},
      {"A": 13, "B": 12, "Z": "poop", "created": "2022/02/01"},
      {"A": 33, "B": 10, "Z": "moop", "created": "2022/02/01"},
      {"A": 0, "B": 5, "Z": "doop", "created": "2022/02/01"},
      {"A": 13, "B": 5, "Z": "poop", "created": "2022/03/01"},
      {"A": 3, "B": 2, "Z": "moop", "created": "2022/03/01"},
      {"A": 20, "B": 7, "Z": "doop", "created": "2022/03/01"}
    ]
  },
  "transform": [
    {"timeUnit": "month", "field": "created", "as": "month"},
    {"fold": ["A", "B"], "as": ["series", "value"]},
    {"calculate": "datum.series + ': ' + datum.Z", "as": "grouping"}
  ],
  "mark": "bar",
  "encoding": {
    "x": {"timeUnit": "month", "field": "created", "type": "ordinal"},
    "y": {"field": "value", "type": "quantitative"},
    "color": {"field": "grouping"}
  }
}

Editor link

Regards,

Daniel





Did I answer your question? Mark my post as a solution!

Proud to be a Super User!


On how to ask a technical question, if you really want an answer (courtesy of SQLBI)




a haaa!  Excellent.  I think this one will work.  I'll give it a go shortly with the real data, but the poops, moops and doops look right in this.

Helpful resources

Announcements
August Power BI Update Carousel

Power BI Monthly Update - August 2025

Check out the August 2025 Power BI update to learn about new features.

August 2025 community update carousel

Fabric Community Update - August 2025

Find out what's new and trending in the Fabric community.

Top Solution Authors