Skip to content
This repository was archived by the owner on Aug 5, 2026. It is now read-only.

WIP Set max character count for byName index in subgraph_domains table #1815

Closed
tk-o wants to merge 4 commits into from
Closed

WIP Set max character count for byName index in subgraph_domains table #1815

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
9cbc064
Enforce maximum value lenght for the `byName` index in `subgraph_doma…
tk-o Mar 24, 2026
b666142
docs(changeset): Updated max character count for the `byName` index d…
tk-o Mar 24, 2026
32f78e0
Update `byName` index to be a partial index
tk-o Mar 24, 2026
cf8a6f3
Fix index creation error
tk-o Mar 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions .changeset/funky-mails-roll.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
---
"@ensnode/ensdb-sdk": patch
---

Updated the max character count (length constraint) for the `byName` partial index defined on the `subgraph_domains` table in the "abstract" ENSIndexer Schema.
20 changes: 18 additions & 2 deletions packages/ensdb-sdk/src/ensindexer-abstract/subgraph.schema.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
import { index, onchainTable, relations } from "ponder";
import { index, onchainTable, relations, sql } from "ponder";
import type { Address } from "viem";

import { monkeypatchCollate } from "../lib/collate";
Expand Down Expand Up @@ -93,7 +93,23 @@ export const subgraph_domain = onchainTable(
expiryDate: t.bigint(),
}),
(t) => ({
byName: index().on(t.name),
/**
* Maximum character length for the `subgraph_domain.name` partial index.

Copilot AI Mar 24, 2026

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The comment refers to subgraph_domain.name, but the actual table name is subgraph_domains (and this index is created on that table). Consider updating the wording to reference subgraph_domains.name to avoid confusion when debugging in Postgres.

Suggested change
* Maximum character length for the `subgraph_domain.name` partial index.
* Maximum character length for the `subgraph_domains.name` partial index.

Copilot uses AI. Check for mistakes.
*
* PostgreSQL B-tree indexes have a maximum size of 8,191 bytes (8KB) per entry.
* Without a limit, index creation fails for entries exceeding this size.
* For example, spam domain records with thousands of characters in `name` column
* cause index creation to fail.
*
* UTF-8 uses 1-4 bytes per character. For the safe maximum of 4 bytes:
* 8,191 ÷ 4 = ~2,000 characters would be the upper bound.
*
* We use 255, which is sufficient for virtually all ENS names in practice.
* This is implemented as a partial index with WHERE length(name) <= 255,
* which excludes oversized spam entries while still allowing normal lookups
* (e.g. where: { name: "foo.eth" }) to use the index.
Comment on lines +109 to +110

Copilot AI Mar 24, 2026

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The doc comment claims normal lookups like where: { name: "foo.eth" } will use this partial index, but in practice the GraphQL layer builds eq(column, value) predicates (parameterized) without length(name) <= 255. Postgres can only use a partial index when the query’s WHERE clause implies the index predicate, so name = $1 alone typically won’t qualify. Consider either (a) adjusting the query builder to add length(name) <= 255 when the provided filter value is <= 255 chars (to preserve semantics), or (b) changing to an expression index approach paired with matching query generation. At minimum, please update this comment to avoid overstating index usage.

Suggested change
* which excludes oversized spam entries while still allowing normal lookups
* (e.g. where: { name: "foo.eth" }) to use the index.
* which excludes oversized spam entries while still indexing typical-length names.
* Note: PostgreSQL can only use this partial index when a query's WHERE clause
* implies the predicate (e.g. includes a matching length constraint), so simple
* equality filters alone (name = $1) may not always qualify for index usage.

Copilot uses AI. Check for mistakes.
*/
byName: index().on(t.name).where(sql`length(${t.name}) <= 255`),
byLabelhash: index().on(t.labelhash),
byParentId: index().on(t.parentId),
byOwnerId: index().on(t.ownerId),
Expand Down
Loading