ClickHouse: MATERIALIZED Column Not Accessible Through VIEW

SkillDatabases & data

Fix HTTP 500 / ClickHouse "column not found" errors when filtering by a MATERIALIZED column through a VIEW that doesn't expose it. Use when: (1) a WHERE clause references alias.column on a view but the column is MATERIALIZED on the underlying table, (2) the query works on the raw table but fails through the view, (3) adding a new filter param to an API causes 500 even though the column exists in the base table. Fix by using a subquery against the base table instead of referencing the column directly on the view. Applies to ClickHouse views over tables with MATERIALIZED or ALIAS columns.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the ClickHouse: MATERIALIZED Column Not Accessible Through VIEW skill

What this skill tells your AI

The instructions your AI receives, as published by divinevideo/divine-mobile in .agents/skills/clickhouse-materialized-column-view-filter/SKILL.md and read by ahel’s review.

Problem

When a ClickHouse VIEW selects specific columns from a table (not SELECT *), any MATERIALIZED or ALIAS columns not explicitly included in the view's SELECT list are invisible to queries through the view. Attempting WHERE v.materialized_col = ? on such a view produces a "column not found" error, which surfaces as an HTTP 500 in API layers.

Context / Trigger Conditions

  • You add a new query filter (e.g., ?platform=vine) that references a column via a view alias
  • The column is defined as String MATERIALIZED ... on the underlying table
  • The VIEW was created with an explicit column list (not SELECT *)
  • The column works fine when querying the base table directly
  • The API returns HTTP 500 with no useful error message to the client
  • Server logs show a ClickHouse "column not found" or similar schema error

Solution

Option A: Subquery (No migration required)

Replace direct column reference with a subquery against the base table:

-- BROKEN: view doesn't expose 'platform'
WHERE v.platform = ?

-- FIXED: subquery against the base table where MATERIALIZED column exists
WHERE v.id IN (
  SELECT id FROM events_deduped
  WHERE platform = ? AND kind IN (34235, 34236)
)

Option B: Migration (Cleaner long-term)

Create a new migration that drops and recreates the view to include the column:

DROP VIEW IF EXISTS nostr.videos;
CREATE VIEW nostr.videos AS
SELECT
    id, pubkey, created_at, kind, content, tags, sig, indexed_at,
    d_tag, title, thumbnail, video_url, author_name, loops,
    platform,  -- ADD THE MATERIALIZED COLUMN
    if(published_at > 0, published_at, toUnixTimestamp(created_at)) AS published_at,
    expiration_at
FROM nostr.events_deduped FINAL
WHERE kind IN (34235, 34236);

Warning: Dropping a view cascades — any dependent views (video_stats, trending_videos, videos_with_loops, etc.) must also be dropped and recreated in the correct dependency order.

Verification

  1. Query the view directly: SELECT platform FROM videos LIMIT 1 — should return data (or empty string for non-vine)
  2. API call with the filter param returns 200 instead of 500
  3. Run full smoke test suite to confirm no regressions

Example (Funnelcake)

The nostr.videos view (migration 000060) selects a fixed column list from events_deduped. The platform column is String MATERIALIZED on events_deduped but not in the view.

PR #85 added v.platform = ? to get_recent_videos_with_events() and get_trending_videos_with_events(), both of which query FROM videos v. This caused HTTP 500 for any request with ?platform=vine.

Fix (PR #86): Changed to subquery approach. Note that videos_with_loops (a different view) DOES include platform — queries through that view (like get_videos_filtered) work fine.

Notes

  • MATERIALIZED columns are physically stored but only accessible if explicitly selected
  • ALIAS columns are computed on read and have the same visibility constraint in views
  • Always check the view definition before adding WHERE conditions on columns
  • The videos_with_loops view includes more columns than videos — consider which view your query is actually using
  • In Funnelcake: videos view = minimal columns; videos_with_loops = full columns including platform

Signals

GitHub stars
264
Forks
55
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
clickhouse-materialized-column-view-filter
Source
github.com/divinevideo/divine-mobile