How to show x-axis months with missing data?

[khustochka]

Hi all,

I have data like this (sample), basically a date, and a species id that was observed:

{"date": "2025-11-03", "quantity": "1", "species_id": 300},
{"date": "2025-11-03", "quantity": "1", "species_id": 296},
{"date": "2025-11-04", "quantity": "1", "species_id": 125},
{"date": "2025-11-04", "quantity": "1", "species_id": 479},
{"date": "2025-11-04", "quantity": "1", "species_id": 413}

My bar chart looks like this (omitting non important stuff):

  "encoding": {
    "x": {
      "axis": {"labelAngle": 0},
      "field": "date",
      "timeUnit": "month",
      "title": "Month",
      "type": "ordinal"
    },
    "y": {"aggregate": "distinct", "field": "date", "title": "Observation days"}
  },
  "mark": "bar",
  "transform": [
    {"filter": "datum.species_id == 275"}
  ]

The result I am getting includes only months when the species was observed:

But I'd like the chart to show all months from January to December. How I can do this? I tried to wrap my head around impute option, but I fail to understand how to apply it to this case. (Also ideally I'd like an option that works for all species, including those that have observations every month.)

Simplified example in Vega editor

Many thanks!

[mattijn]

Thanks for raising! It would be great if you can share a minimal reproducible example that people can open in the Vega editor.

Based on what you shared I see that you are using ”type”: “ordinal” on the x encoding channel, where I would expect an ”type”: “temporal”.

[khustochka]

Thanks!

I have added the link to Vega editor example to the post.

Changing ordinal to temporal produces the same result with missing months for which there are no data. It just looks a bit different.

[khustochka]

Yay, I think I have fixed it. I used ChatGPT (sorry) and although its suggestions did not work as is, I have figured out in what direction to look and was able to fix it. The trick is to do aggregations in transform, give them names, and now I can use impute.

  {"calculate": "month(toDate(datum.date))", "as": "month"},
  {
    "aggregate": [{"op": "distinct", "field": "date", "as": "distinct_days"}],
    "groupby": ["month"]
  },
  {
    "impute": "distinct_days",
    "key": "month",
    "keyvals": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11],
    "value": 0
  }

Solution