postgresql/prefer-keyword-case

Why this matters

Operates on the SQL keyword tokens emitted by the parser; identifiers, string literals, and comments are left alone. The `types` option separately controls how built-in type names (`int`, `text`, `varchar`) are cased — defaulting to `skip` so user-defined type names like `ulid` (which the rule cannot touch) don't end up case-mismatched against built-ins.

Examples

select id, name from users where active = true limit 10;

SELECT id, name FROM users WHERE active = TRUE LIMIT 10;

Configure it

// eslint.config.js
import postgresql from "eslint-plugin-postgresql";

export default [
  {
    files: ["**/*.sql"],
    languageOptions: {
      parser: postgresql.configs.recommended.languageOptions.parser,
    },
    plugins: { postgresql },
    rules: {
      "postgresql/prefer-keyword-case": [
        "warn",
        {
          case: "upper",
          types: "skip",
        },
      ],
    },
  },
];

Options

case "upper" | "lower" default: "upper"

Target case for SQL keywords.

types "upper" | "lower" | "skip" default: "skip"

How to case built-in type names. `skip` leaves them as written so they don't case-mismatch against user-defined types the rule cannot touch.