Hello, fellow developers! I keep forgetting how to query GROQ, so here’s a cheat sheet covering only the essentials you’ll actually use.

Basic fetch

*[_type == "post"]

Everything starts here: * (all documents), filtered by _type.

Filter

*[_type == "post" && !(_id in path("drafts.**"))]

&& for chaining conditions
! for negation
Drafts live under "drafts." path—exclude them explicitly

Sort

*[_type == "post"]
  | order(publishedAt desc)

Pipe | passes results forward
asc or desc
You can chain multiple sorts if needed

Limit / Slice

*[_type == "post"]
  | order(publishedAt desc)
  [0...5]

[start...end]
Zero-based index
Great for pagination or “latest posts”

Select fields

*[_type == "post"]{
  title,
  publishedAt
}

Projection {} trims the payload
Only return what you need

Rename fields

*[_type == "post"]{
  "date": publishedAt
}

"alias": originalField
Useful for frontend consistency

Slug gotcha

*[_type == "post"]{
  "slug": slug.current
}

slug is an object
Always access .current
Otherwise you’ll get {_type: "slug", current: "..."}

Cleaning arrays

*[_type == "post"]{
  "tags": tags[]
}

tags[] extracts raw values
Without [], you might get richer objects depending on schema

Nested projection

*[_type == "post"]{
  metadata[]{
    key,
    value
  }
}

Each item is an object
You must project inside the array
Think: map over array, pick fields

Dereferencing

*[_type == "post"]{
  title,
  author->{
    name,
    avatar
  }
}

-> follows references
Works like a lightweight join

Conditional fields

*[_type == "post"]{
  title,
  "isFeatured": featured == true
}

Inline computed fields
Can simplify frontend logic

Fallback values

*[_type == "post"]{
  title,
  "excerpt": coalesce(excerpt, "No summary")
}

coalesce(a, b) returns first non-null
Handy for incomplete content

Count

count(*[_type == "post"])

Returns a number
Useful for pagination metadata

Match / Search

*[_type == "post" && title match "sanity*"]

Basic text search
* acts like a wildcard

Combine it all

*[_type == "post" && defined(slug.current)]
  | order(publishedAt desc)
  [0...5]{
    title,
    "slug": slug.current,
    "tags": tags[],
    metadata[]{
      key,
      value
    }
  }

Mental model

Filter → Sort → Slice → Project
Arrays need []
Objects need {}
References need ->
Slugs need .current

That’s most of what you’ll reach for day-to-day.

Sanity GROQ query cheatsheet

Table of contents

Table of contents