Skip to main content

Query Parsing

Query strings are available through ctx.query. It is a record of string | string[]; repeated keys become arrays.

export async function GET(ctx: BurgerContext) {
  const { tag, page } = ctx.query;
  // ...
}

Why not new URL(ctx.url)?​

new URL(ctx.url) builds a URL object and a URLSearchParams instance on every call, using memory even when you only need a value or two. ctx.query instead:

  • Parses lazily, only when you first read it, and caches the result.
  • Uses a fast parser that matches URLSearchParams semantics, including + decoding to a space and lenient handling of malformed escapes.
  • Avoids allocating a URL object on every request.

So ctx.query is both simpler to read and cheaper when the query is unused.

Behavior​

  • ctx.query is parsed once on first access, then cached. A request that never reads it pays nothing.
  • Repeated keys become arrays, in order.
  • Malformed percent-encoding is preserved and never throws.
// Same result, without the extra memory use:
const tag = ctx.query.tag; // instead of new URL(ctx.url).searchParams.get("tag")