Query index: background, definition syntax, troubleshooting

Sources: Indexing, Indexing reference, Spreadsheets and JSON, Sitemaps, Limits and the Admin API, read in November 2026. Edge Delivery Services changes: check a detail there before you rely on it.

How the index works

• An index lists the published pages of a part of the site. A page enters or updates when it is published, by Sidekick or by the API alike. It leaves when it is unpublished. Previewing changes nothing.
• The definition is managed in the Index Admin tool or with the Admin API (/config/<org>/sites/<site>/content/query.yaml). After a change to the properties, Reindex in the tool reindexes the whole site.
• A site can have several indexes. Each is served at the target of its definition. /query-index.json is only the default name.
• With SharePoint or Google Drive as the content source, the index can also live in a spreadsheet. The indexer writes the sheet raw_index. A sheet named helix-default with a FILTER formula over raw_index can leave out pages, for example those with noindex. This is not available with Document Authoring, AEM authoring or another bring-your-own-markup source: there, filter in the consumer by the robots property.

The response

{ "total": 1472, "offset": 0, "limit": 1000, "data": [{ "path": "/…" }], "columns": ["path", "title"], ":type": "sheet" }

• Without a limit parameter, at most 1000 entries are returned. Page with ?offset=<n>&limit=<n>.

Limits

Definition syntax

indices:
  site:
    include:
      - /**
    exclude:
      - /drafts/**
    target: /query-index.json
    properties:
      title:
        select: head > meta[property="og:title"]
        value: attribute(el, "content")
      description:
        select: head > meta[name="description"]
        value: attribute(el, "content")
      image:
        select: head > meta[property="og:image"]
        value: attribute(el, "content")
      robots:
        select: head > meta[name="robots"]
        value: attribute(el, "content")
      lastModified:
        select: none
        value: parseTimestamp(headers["last-modified"], "ddd, DD MMM YYYY hh:mm:ss GMT")
  blog:
    include:
      - /blog/**
    target: /blog/query-index.json
    properties:
      title:
        select: head > meta[property="og:title"]
        value: attribute(el, "content")
      author:
        select: head > meta[name="author"]
        value: attribute(el, "content")
      tags:
        select: head > meta[property="article:tag"]
        values: attribute(el, "content")

• include and exclude decide which pages are indexed. /** matches everything under a prefix and the prefix itself. Without the section, every page is indexed.
• select is a CSS selector on the delivered HTML. It returns all matches; selectFirst returns the first. Not every selector is supported: the indexer uses hast-util-select.
• value gives a string, values an array.
• Functions: attribute(el, name), textContent(el), innerHTML(el), match(el, re), words(el, start, end), characters(el, start, end), replace(el, substr, newSubstr), replaceAll(el, substr, newSubstr), parseTimestamp(el, format), dateValue(el, format). Variables: el, path, headers[name].
• path needs no property: every entry has it.

The Index Admin tool offers title, description, image, author, date, category, tags, robots and lastModified. Those whose source is not the meta tag of the same name:

Troubleshooting