identifierTooLong¶

Group correctness · Recommended · Safe fix · Since Unreleased
Diagnostic lint/identifierTooLong

Sources: inspired by squawk/identifier-too-long

Description¶

Identifiers longer than 63 bytes are silently truncated.

Postgres truncates longer identifiers to 63 bytes and only raises a notice, so the names you write aren't the names Postgres uses. Two long names that share their first 63 bytes even end up as the same name.

Examples¶

Invalid¶

create table customer_subscription_billing_cycle_adjustments_history_archives_v2 (id bigint);

code-block.sql:1:14 lint/identifierTooLong  FIXABLE  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  ! Identifier customer_subscription_billing_cycle_adjustments_history_archives_v2 is longer than 63 bytes.

  > 1 │ create table customer_subscription_billing_cycle_adjustments_history_archives_v2 (id bigint);
      │              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    2 │ 

  i Postgres truncates it to `customer_subscription_billing_cycle_adjustments_history_archive`.

  i Use a shorter name.

Valid¶

create table subscription_adjustments_archive (id bigint);

How to configure¶

{
  "linter": {
    "rules": {
      "identifierTooLong": "error"
    }
  }
}

How to suppress¶

Suppress this diagnostic with a comment:

-- pgls-ignore identifierTooLong