--- title: Transform --- The `transform` function of a collection can be used to perform various tasks on the document before it is saved. It receives two arguments: the document, which matches the schema's shape, and a context object. The result of the `transform` function defines the TypeScript type of the document and should return an object that can be serialized to JSON. The function can be synchronous or asynchronous. ## Content In Content Collections, the content of a document is just a string. One common use case for the `transform` function is to transform this content, such as converting markdown to HTML. While the content can be used directly with libraries like [react-markdown](https://github.com/remarkjs/react-markdown), transforming the content during build time is often beneficial. Various libraries available in the Node.js ecosystem can be used for transformation, with the only requirement being that the result is [serializable](/docs/serialization). However, there are specific helper packages designed for Content Collections: - [@content-collections/markdown](/docs/content/markdown) for converting markdown to HTML - [@content-collections/mdx](/docs/content/mdx) for converting MDX to JSX Another option is to use Content Collections solely for metadata (e.g., frontmatter) while employing a different tool to render the content (e.g., `next/mdx`). This approach is particularly beneficial for larger sites using `mdx`. If the bundler that compiles the rest of the application also includes the `mdx` content files, it can optimize the resulting bundle. This way, imports in `mdx` files are resolved by the bundler and function like those in the rest of the application. Using an external bundler, as we do with `@content-collections/mdx`, may lead to larger bundles because the external bundler must re-bundle every used component, even if it is already utilized elsewhere in the application. To implement this approach, the `frontmatter-only` parser can be used alongside [Static Import](#static-imports). For a complete sample with Next.js, refer to the [next-mdx-static-import sample](/samples/next-mdx-static-import). ## Caching The `transform` function is invoked for every document during each build or document change. While this behavior may change in the future, it currently holds true. Therefore, ensuring fast transformations is crucial. For slower operations like markdown to HTML conversion, the context object provides a `cache` function to cache costly operations. Example: ```ts transform: async (doc, { cache }) => { const html = await cache(doc.content, async (content) => { return markdownToHtml(content); }); return { ...doc, html, }; }; ``` In this example, the `markdownToHtml` function is only called if the content has changed. **Note**: Caching the compilation steps of `@content-collections/markdown` or `@content-collections/mdx` is unnecessary as they already utilize the same caching mechanism. ## Access sibling documents Since version 0.7.0, it is possible to access other documents of the same collection by using the `documents` function of the `collection` object, which is part of the `context` object. The function is asynchronous, requires no parameters, and returns an array of all documents of the collection. The documents are not transformed; they have the shape as defined in the schema of the collection. Example: ```ts const posts = defineCollection({ // ... transform: async (doc, { collection }) => { const docs = await collection.documents(); const idx = docs.findIndex((d) => doc._meta.filePath === d._meta.filePath); return { ...doc, prev: idx > 0 ? docs[idx - 1] : null, next: idx < docs.length - 1 ? docs[idx + 1] : null, }; }, }); ``` ## Access other collections The `transform` function can access other collections using the `documents` function of the `context` object. The function requires a collection reference as parameter and returns an array of documents for that collection. But keep in mind the returned document are not transformed, they have the shape as defined in the schema of the referenced collection. Example: ```ts const authors = defineCollection({ // ... schema: z.object({ ref: z.string(), displayName: z.string() }), }); const posts = defineCollection({ // ... transform: async (doc, { documents }) => { const author = await documents(authors).find( (a) => a.ref === doc.author ); return { ...doc, author: author.displayName }; }, ``` For a complete example have a look at the [Join collections](#join-collections) example. It is not possible to access documents of the same collection with the `documents` function. Use the `collection.documents` function instead. Please refer to [Access sibling documents](#access-sibling-documents) for more information. ## Static Imports Since version 0.8.0, the `transform` function can return a reference to a static import. This reference can be used to import a JavaScript module, which then becomes part of the generated document. Example: ```ts transform: async (doc) => { const component = createDefaultImport(`@/components/${doc.component}`); return { ...doc, component, }; }, ``` The static import is very useful when working with MDX files, and the application's bundler allows for their import. Example: ```ts transform: async (doc) => { const mdxContent = createDefaultImport(`@/content/posts/${doc._meta.filePath}`); return { ...doc, mdxContent, }; }, ``` The `mdxContent` can the be used in the application: ```tsx const Post = ({ post }: Props) => { const MDXContent = post.mdxContent; return ; }; ``` To ensure type safety, a generic for the imported module must be passed to the `createDefaultImport`. ## Skipping documents Since version 0.11.0, the `transform` function can skip a document from being added to the collection. This feature is useful for filtering out documents that do not meet specific criteria or for preventing certain documents from being processed. To skip a document, simply return the result of the `skip` function from the context object with an optional reason, for example: ```ts transform: (doc, { skip }) => { if (doc.draft) { return skip("document is a draft"); } return doc; }); ``` ## Examples Here are some common use cases of the `transform` function: ### Generate a slug ```ts transform: (doc) => { return { ...doc, slug: doc.title.toLowerCase().replace(/ /g, "-"), }; }; ``` ### Fetch external data ```ts transform: async (author, { cache }) => { const bio: string = await cache( `https://api.authors.com/bio/${author.username}`, async (url) => { const response = await fetch(url); const user = await response.json(); return user.bio; }, ); return { ...user, bio, }; }; ``` ### Get last modification date from Git ```ts transform: async (doc, { cache }) => { const lastModified = await cache(doc._meta.filePath, async (filePath) => { const { stdout } = await exec(`git log -1 --format=%ai -- ${filePath}`); if (stdout) { return new Date(stdout.trim()).toISOString(); } return new Date().toISOString(); }); return { ...doc, lastModified, }; }; ``` ### Join collections ```ts import { defineCollection, defineConfig } from "@content-collections/core"; import { z } from "zod"; const authors = defineCollection({ name: "authors", directory: "content/authors", include: "*.md", schema: z.object({ ref: z.string(), displayName: z.string(), email: z.string().email(), content: z.string(), }), }); const posts = defineCollection({ name: "posts", directory: "content/posts", include: "*.md", schema: z.object({ title: z.string(), author: z.string(), content: z.string(), }), transform: async (document, context) => { const author = await context .documents(authors) .find((a) => a.ref === document.author); return { ...document, author, }; }, }); export default defineConfig({ content: [authors, posts], }); ```