:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
import { PART_OF_WHOLE_DISPLAY_TYPES } from 'lib/constants'
import { getTableDisplayOptions } from '~/queries/nodes/DataVisualization/Components/TableDisplay'
import {
applyVisualizationType,
columnsFromResponse,
getAutoVisualizationType,
} from '~/queries/nodes/DataVisualization/dataVisualizationLogic'
import { sqlVisualizationDisabledReason } from '~/queries/nodes/DataVisualization/sqlVisualizationSupport'
import { DataVisualizationNode, NodeKind } from '~/queries/schema/schema-general'
import { ChartDisplayType } from '~/types'
describe('dashboard SQL visualization support', () => {
const responses = {
'date and numeric': {
columns: ['day', 'total'],
types: [
['day', 'DateTime'],
['total', 'UInt64'],
],
result: [
['2026-01-01', 1],
['2026-01-02', 2],
],
},
'two strings and a numeric, which Auto resolves to a 2d heatmap': {
columns: ['country', 'browser', 'hits'],
types: [
['country', 'String'],
['browser', 'String'],
['hits', 'UInt64'],
],
result: [['NL', 'Firefox', 3]],
},
'all numeric, which the editor plots by promoting the first column to the x axis': {
columns: ['users', 'events'],
types: [
['users', 'UInt64'],
['events', 'UInt64'],
],
result: [[1, 2]],
},
'all string, so nothing is left to plot': {
columns: ['country', 'browser'],
types: [
['country', 'String'],
['browser', 'String'],
],
result: [['NL', 'Firefox']],
},
'a single numeric column': {
columns: ['total'],
types: [['total', 'UInt64']],
result: [[7]],
},
}
const baseQuery = {
kind: NodeKind.DataVisualizationNode,
source: { kind: NodeKind.HogQLQuery, query: 'select 1' },
} as DataVisualizationNode
it.each(Object.entries(responses))(
'every type the card leaves enabled saves a query it can draw: %s',
(_label, response) => {
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
const numericalColumns = columns.filter((column) => column.type.isNumerical)
const options = getTableDisplayOptions(columns, numericalColumns, autoVisualizationType, (displayType) =>
sqlVisualizationDisabledReason(
displayType,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
)
const enabled = options
.flatMap((group: any) => (Array.isArray(group.options) ? group.options : []))
.filter((option: any) => !option.disabledReason)
.map((option: any) => option.value as ChartDisplayType)
expect(enabled.length).toBeGreaterThan(0)
const chartTypes = enabled.filter((displayType) => {
const resolved =
displayType === ChartDisplayType.Auto ? autoVisualizationType : (displayType as ChartDisplayType)
return ![ChartDisplayType.ActionsTable, ChartDisplayType.BoldNumber].includes(resolved)
})
for (const displayType of chartTypes) {
const saved = applyVisualizationType(baseQuery, displayType, columns, response.result.length)
// A part-of-whole chart draws one part per value column, so it needs no X-axis.
const needsXAxis = !PART_OF_WHOLE_DISPLAY_TYPES.includes(displayType)
expect(!needsXAxis || typeof saved.chartSettings?.xAxis?.column === 'string').toBe(true)
expect(saved.chartSettings?.yAxis?.length ?? 0).toBeGreaterThan(0)
}
}
)
it.each([ChartDisplayType.ActionsPie, ChartDisplayType.ActionsDonut, ChartDisplayType.ActionsProportionBar])(
'the SQL editor picker disables %s when the results have no numeric column',
(displayType) => {
const columns = columnsFromResponse({
columns: ['country', 'browser'],
types: [
['country', 'String'],
['browser', 'String'],
],
result: [['US', 'Chrome']],
})
const option = getTableDisplayOptions(columns, [], ChartDisplayType.ActionsTable)
.flatMap((group: any) => (Array.isArray(group.options) ? group.options : []))
.find((candidate: any) => candidate.value === displayType)
expect(option?.disabledReason).toBe('Requires at least one numeric column')
}
)
it.each([responses['date and numeric'], responses['a single numeric column']])(
'sets up a dashboard Metric from numeric results',
(response) => {
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
expect(
sqlVisualizationDisabledReason(
ChartDisplayType.Metric,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
).toBeUndefined()
const saved = applyVisualizationType(baseQuery, ChartDisplayType.Metric, columns, response.result.length)
expect(saved.chartSettings?.yAxis).toHaveLength(1)
}
)
it.each([responses['date and numeric'], responses['a single numeric column']])(
'sets up a dashboard proportion bar from numeric results, like Metric',
(response) => {
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
expect(
sqlVisualizationDisabledReason(
ChartDisplayType.ActionsProportionBar,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
).toBeUndefined()
}
)
it.each([ChartDisplayType.ActionsProportionBar, ChartDisplayType.ActionsPie, ChartDisplayType.ActionsDonut])(
'allows a dashboard %s from numeric-only results, with one part per column',
(displayType) => {
const response =
responses['all numeric, which the editor plots by promoting the first column to the x axis']
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
expect(
sqlVisualizationDisabledReason(
displayType,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
).toBeUndefined()
}
)
it('still requires a numeric column for a proportion bar', () => {
const response = responses['all string, so nothing is left to plot']
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
expect(
sqlVisualizationDisabledReason(
ChartDisplayType.ActionsProportionBar,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
).toEqual('This insight has no numeric column to plot')
})
it('sets up a dashboard horizontal bar chart from category and numeric columns', () => {
const response = responses['date and numeric']
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
expect(
sqlVisualizationDisabledReason(
ChartDisplayType.ActionsBarValue,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
).toBeUndefined()
const saved = applyVisualizationType(
baseQuery,
ChartDisplayType.ActionsBarValue,
columns,
response.result.length
)
expect(saved.chartSettings!.xAxis!.column).toBe('day')
expect(saved.chartSettings!.yAxis!.map((series) => series.column)).toEqual(['total'])
})
it('does not offer Auto when it resolves to a type the card cannot set up', () => {
const response = responses['two strings and a numeric, which Auto resolves to a 2d heatmap']
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
expect(autoVisualizationType).toEqual(ChartDisplayType.TwoDimensionalHeatmap)
expect(
sqlVisualizationDisabledReason(
ChartDisplayType.Auto,
baseQuery,
columns,
response.result.length,
autoVisualizationType
)
).toContain('Open the insight')
})
it('does not treat stale saved axes as making a chart type available', () => {
const response = responses['all string, so nothing is left to plot']
const columns = columnsFromResponse(response)
const autoVisualizationType = getAutoVisualizationType(columns, response.result.length)
const staleAxes = {
...baseQuery,
chartSettings: { xAxis: { column: 'country' }, yAxis: [{ column: 'browser' }] },
} as DataVisualizationNode
for (const candidate of [baseQuery, staleAxes]) {
expect(
sqlVisualizationDisabledReason(
ChartDisplayType.ActionsBar,
candidate,
columns,
response.result.length,
autoVisualizationType
)
).toEqual('This insight has no numeric column to plot')
}
})
it('preserves saved axes while response columns are unavailable', () => {
const query = {
...baseQuery,
chartSettings: {
xAxis: { column: 'day' },
yAxis: [
{
column: 'total',
settings: { formatting: { prefix: '$', suffix: '' } },
},
],
},
} as DataVisualizationNode
expect(applyVisualizationType(query, ChartDisplayType.ActionsBar, [], 0).chartSettings).toEqual(
query.chartSettings
)
})
})