: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 { HogFunctionTemplate } from '~/cdp/types'
export const template: HogFunctionTemplate = {
free: true,
status: 'stable',
type: 'transformation',
id: 'template-downsampling',
name: 'Downsampling',
description:
'Keeps only a percentage of events to reduce volume. Stable sampling keeps the same distinct IDs as the percentage grows; random sampling decides per event.',
icon_url: 'https://res.cloudinary.com/dmukukwp6/image/upload/q_auto,f_auto/builder_hog_01_955c082cad.png',
category: ['Custom'],
code_language: 'hog',
code: `
// Only sample the events named here; empty means sample every event.
let applies := true
if (not empty(trim(inputs.triggeringEvents))) {
applies := false
for (let name in splitByString(',', inputs.triggeringEvents)) {
if (trim(name) == event.event) {
applies := true
}
}
}
if (not applies) {
return event
}
// Handle the boundaries directly so 0 drops every event and 100 keeps every event in both modes,
// and an out-of-range percentage cannot keep or drop more than intended.
if (inputs.percentage <= 0) {
return null
}
if (inputs.percentage >= 100) {
return event
}
let keep := true
if (inputs.randomSampling) {
keep := (randomFloat() * 100) <= inputs.percentage
} else {
// Map the distinct ID to a stable value in [0, 1] from its sha256 hash. A distinct ID kept
// at a lower percentage stays kept as the percentage grows.
let hex := sha256Hex(event.distinct_id)
let digits := '0123456789abcdef'
let acc := 0.0
let maxAcc := 0.0
let i := 1
while (i <= 13) {
let d := position(digits, substring(hex, i, 1)) - 1
if (d < 0) {
d := 0
}
acc := acc * 16 + d
maxAcc := maxAcc * 16 + 15
i := i + 1
}
keep := (acc / maxAcc) <= (inputs.percentage / 100)
}
if (keep) {
return event
}
return null
`,
inputs_schema: [
{
key: 'percentage',
type: 'number',
label: 'Percentage of events to keep',
description: 'A number between 0 and 100. 100 keeps every event, 0 drops every event.',
default: 100,
required: true,
},
{
key: 'randomSampling',
type: 'boolean',
label: 'Use random sampling',
description:
'When off, sampling is stable per distinct ID, so raising the percentage only adds events. When on, each event is decided independently.',
default: false,
required: false,
},
{
key: 'triggeringEvents',
type: 'string',
label: 'Events to sample',
description: 'Comma-separated event names to sample. Leave empty to sample every event.',
default: '',
required: false,
},
],
}