Astro.glob versus getCollection


Astro provides two ways to collect document content:

  1. Astro.glob
  2. getCollection from Astro:content

Confoundingly, each represents content data differently.


The code below Astro.glob fetches the posts content collection. The argument passed to Astro.glob must a be string literal.

const posts = await Astro.glob("../content/posts/*.mdx");

The code above fetches .mdx files, but you could get both Markdown and mdx files with the *.md? file selector.

Astro.glob returns this object as one of the elements of the content data:

  frontmatter: [Getter],
  getHeadings: [Getter],
  __usesAstroImage: [Getter],
  url: [Getter],
  file: [Getter],
  Content: [Getter],
  default: [Function: Content] {
    moduleId: 'C:/Users/thumb/Documents/Projects/astro-4/blog-again/src/content/posts/_template.mdx',
    [Symbol(mdx-component)]: true,
    [Symbol(astro.needsHeadRendering)]: false
  [Symbol(Symbol.toStringTag)]: 'Module'


The code below fetches a content collection with getCollection. It’s a little more flexible than Astro.glob because it isn’t location-specific, it fetches the content collection requested.

import { getCollection } from "astro:content";
const posts = await getCollection("posts");

getCollection returns this as one of the elements of the content data:

  id: 'astro-notes.mdx',
  slug: 'astro-notes',
  body: '[body]',
  collection: 'posts',
  data: {
    title: 'Astro notes',
    description: 'Astro notes',
    tags: [ 'astro' ],
    date_published: 2023-04-16T00:00:00.000Z,
    date_added: 2023-04-16T00:00:00.000Z,
    date_updated: 2023-04-16T00:00:00.000Z,
    draft: false
  render: [AsyncFunction: render]

Because these two methods return distinct object structures, helpers like content filters and queries must be object-specific. Both are wrappers around Vite’s glob function so their performance is similar.

Although getCollection requires an import, I use it exclusively to fetch content collections. The helper function makes it easy to get all the data needed from the a content collection member to list it.

export function getContentData(obj) {
  return {
    date_published: || "N/A",
    slug: `/${obj.collection}/${obj.slug}`,

My convention is that if data_published is null, then the document is a draft otherwise it’s been published. If it is a draft document getContentData returns “N/A.”

The example below shows an Astro component that emits the HTML needed to the content collection member in a list. It receives a single post as a post as a property and then uses getContentData to get the values needed to list the document.

import { getContentData } from "$scripts/utils.js";

const { post } = Astro.props;
const { title, description, date_published, slug } = getContentData(post);

<div class="post-detail">
  <a href={`${slug}`}>
    <div class="post-description">
      {date_published == "N/A" && <span class="draft-post-tag">Draft</span>}
      {date_published != "N/A" && `${formatDate(date_published)}`}