Managing Path Cleaning Rules

Path cleaning rules normalize $pathname and $entry_pathname so that pages sharing the same template (/users/123/profile, /users/456/profile, …) collapse into one row (/users/<id>/profile) in Web analytics tiles, Paths insights, and any HogQL query that calls apply_path_cleaning. They are the right answer when a breakdown is fragmented across thousands of near-identical URLs.

This skill teaches you how to:

• recognize when path cleaning is the right tool
• inspect real paths to find what needs cleaning
• write regex + alias rules in re2 syntax with the project’s placeholder convention
• test rules before saving them
• order rules so specific patterns aren’t swallowed by generic ones
• apply the rules via MCP

Data model

Team.path_cleaning_filters is a JSON list of PathCleaningFilter objects:

{
  "regex": "/users/\\d+/profile",
  "alias": "/users/<id>/profile",
  "order": 0
}

• regex — a re2 (opens in a new tab) pattern. No need to escape /. Anchor with ^ / $ when you mean it.
• alias — the replacement. Default to angle-bracket placeholders (<id>, <slug>, <uuid>, <date>) so the cleaned path stays human-readable. Backreferences to the regex’s capture groups also work, but they cost more to evaluate — see the backreference pitfall below before reaching for one.
• order — integer. Rules apply sequentially in order ascending, each rule’s output feeds the next.

Application is replaceRegexpAll(pathname, regex, alias) per rule, chained.

Workflow

1. Confirm path cleaning is the right move

Ask yourself: is the user complaining about cardinality (too many distinct paths in a chart), or do they want a per-URL drill-down? Path cleaning is for the former. If they want per-URL data, suggest a property filter on $pathname instead.

2. Inspect the real paths

Don’t guess at patterns — query them. With the execute-sql MCP tool:

SELECT properties.$pathname AS path, count() AS views
FROM events
WHERE event = '$pageview'
  AND timestamp > now() - INTERVAL 7 DAY
GROUP BY path
ORDER BY views DESC
LIMIT 200

Scan the result for:

• numeric IDs: /users/123, /orders/4242
• UUIDs: /sessions/8f3c1a3b-…
• slugs: /posts/why-i-love-posthog
• dates: /archive/2024-09-12
• locales: /en-US/, /fr-FR/
• pagination: ?page=3, /page/3/

3. Draft regex + alias

Anchoring rules of thumb:

• start the regex with ^ only when the segment must be at the beginning of the path
• end with $ to keep a generic rule (e.g. \d+$) from matching mid-path segments

4. Test before saving

Three options, pick one:

• Settings page tester: /settings/project#path_cleaning has a built-in “test path” input that replays the full ordered chain.

• Project HogQL (via execute-sql):

  sqlCopy

  SELECT replaceRegexpAll('/users/42/profile', '/users/\d+/profile', '/users/<id>/profile')

  Chain replaceRegexpAll calls in the same order the rules will run if you want to verify multi-rule interaction.

• Built-in AI helper: there is already an AiRegexHelper modal accessible from the rule editor (Help me with Regex button) that turns natural language into a regex. Suggest it to the user when they say “I don’t know regex” — but always validate the output against real paths via the tester.

5. Order rules from most-specific to most-general

Sequential application means a generic rule placed first will swallow everything that should have hit a specific rule.

order=0  /users/me/profile        →  /users/me/profile     (specific, runs first)
order=1  /users/\d+/profile       →  /users/<id>/profile
order=2  /users/[a-z0-9-]+        →  /users/<slug>          (catch-all, runs last)

If /users/[a-z0-9-]+ ran first it would also match /users/me/profile and make the more specific rule unreachable.

6. Apply via MCP

Prefer the path-cleaning-rules-update tool. It reads the current rules, applies granular operations (append, insert, replace, remove, reorder), auto-numbers order, and — unless you pass confirm: true — returns a preview of the resulting rules without saving. Pass sample_paths to see how the resulting set rewrites real paths.

First call it without confirm to get the preview, surface that to the user, then re-send the same call with "confirm": true to save:

{
  "operations": [{ "action": "append", "alias": "/users/<id>/profile", "regex": "/users/\\d+/profile" }],
  "sample_paths": ["/users/123/profile", "/users/me/profile"]
}

Because the tool does the read-modify-write for you, you don’t have to fetch the full list, renumber order, or risk clobbering existing rules — the common failure mode when editing path_cleaning_filters directly.

If you must fall back to project-settings-update (whole-list overwrite), always read the existing rules first and merge — overwriting silently destroys whatever the team has already configured.

Where the rules apply

When the user (or a HogQL query) opts in:

• Web analytics: the Path cleaning toggle in the page header
• Paths insights: the path cleaning toggle in the insight filters
• HogQL: any query that calls apply_path_cleaning(path_expr, team)

The rules are stored once per project — they are not insight-scoped.

Common pitfalls

• Prefer a placeholder alias over a backreference — a rule that substitutes a capture group costs roughly 3x more per row than the same rule with a fixed alias, because re2 has to track submatch boundaries instead of answering match/no-match on its fast path. Reach for one only when it collapses several near-identical rules into one, which is a clear net win since cleaning evaluates every rule against every row. Keep the regex ^-anchored either way: replaceRegexpAll keeps scanning for further matches, so an unanchored pattern is much more expensive than an anchored one.
• Backreferences in alias need double-escaping — ClickHouse’s replaceRegexpAll supports \0 (whole match) and \1–\9 (capture groups). In a JSON field or SQL string literal the backslash must be doubled, so use \\1 in path_cleaning_filters / HogQL to get the \1 backreference at the ClickHouse layer.
• Forgetting $ — \d+ without an end anchor matches every numeric run in any path, so /blog/2024-09-12/post becomes /blog/<num>-<num>-<num>/post when you only meant to match the year segment. Use \d+$ or \d+(/|$) depending on intent.
• Escaping / — re2 does not require it. \/ works but adds noise.
• Case sensitivity — re2 is case-sensitive by default. Use (?i) at the start of the pattern for case-insensitive matching, e.g. (?i)/users/\d+.
• Replacing the whole list — the raw path_cleaning_filters field is overwrite, not append. Use path-cleaning-rules-update (which does the read-modify-write for you) instead of hand-editing the field; if you must edit it directly, always start from the current list.
• Rules apply globally — adding a rule can change historical numbers in every Web analytics / Paths chart that has cleaning enabled. Warn the user before applying anything destructive.