Filters and ordering

Filters

Every list, count, connection and deleteMany takes a where argument. Each field of the type has a filter input that depends on its type:

Filter	Operators	Used for
IntFilter, FloatFilter	eq ne gt gte lt lte in isNull	Numbers
DateTimeFilter	eq ne gt gte lt lte isNull	Dates and timestamps
StringFilter	eq ne in contains isNull	Text, and bigint/numeric columns
BooleanFilter	eq ne isNull	Booleans
IDFilter	eq ne in isNull	MongoDB ids
NullFilter	isNull	JSON values and arrays

contains is a case-insensitive substring match. isNull: true matches rows where the field is null (or, on MongoDB, missing).

All conditions on one field must hold, and so must the conditions on different fields:

{
  blogPosts(where: { rating: { gte: 3, lt: 5 }, published: { eq: true } }) {
    title
  }
}

AND, OR and NOT

{
  blogPosts(
    where: {
      OR: [{ title: { contains: "graphql" } }, { rating: { gte: 4 } }]
      NOT: { body: { isNull: true } }
    }
  ) {
    title
  }
}

Filtering by a relation

A relation to one row filters on that row's fields:

{
  blogPosts(where: { author: { fullName: { contains: "ada" } } }) { title }
}

A relation to many rows takes some, none or every:

{
  authors(where: { blogPosts: { some: { published: { eq: true } } } }) { fullName }
}

every is true for a row with no related rows at all. On SQL databases these compile to EXISTS subqueries; on MongoDB the related ids are read first.

Ordering

orderBy takes a list; earlier entries take precedence:

{
  blogPosts(orderBy: [{ published: DESC }, { rating: DESC }]) { title }
}

Each entry names one field with ASC or DESC. JSON fields and arrays can't be ordered by. Without orderBy, rows come sorted by primary key, so pages are stable.