@neo4j/graphql@5.12.3

Patch Changes

• #6294 8cf40d9 Thanks @MacondoExpress! - Fix bug that causes 1:1 non-nullable relationship filters to produce invalid Cypher.

@neo4j/graphql-ogm@5.12.3

Patch Changes

• Updated dependencies [8cf40d9]:
  • @neo4j/graphql@5.12.3

@neo4j/graphql@6.6.2

Patch Changes

• #6274 0d94cc4 Thanks @angrykoala! - Backport fix #6243: Handles Neo4j error "52N29" on CDC polling by refreshing the cursor. This error could be triggered in some cases by an outdated cursor.

@neo4j/graphql@5.12.2

Patch Changes

• #6275 8a993a8 Thanks @angrykoala! - Backport fix #6243: Handles Neo4j error "52N29" on CDC polling by refreshing the cursor. This error could be triggered in some cases by an outdated cursor.

@neo4j/graphql-ogm@5.12.2

Patch Changes

• Updated dependencies [8a993a8]:
  • @neo4j/graphql@5.12.2

@neo4j/graphql@7.1.1

Patch Changes

• #6243 8d4c5c6 Thanks @angrykoala! - Handles Neo4j error "52N29" on CDC polling by refreshing the cursor. This error could be triggered in some cases by an outdated cursor.

@neo4j/graphql@7.1.0

Minor Changes

• #6199 4404026 Thanks @angrykoala! - Add support for @cypher directive of fields targeting types that do not use the @node directive. For example:

  type Movie @node {
      title: String
      id: String!
      link: Link!
          @cypher(
              statement: """
              MATCH(l:${Link})
              WHERE l.movieId=this.id
              RETURN l {.name, .url} as link
              """
              columnName: "link"
          )
  }
  
  type Link {
      movieId: String!
      url: String!
      name: String!
  }

Patch Changes

• #6223 de2dcea Thanks @angrykoala! - Fix bug with @populatedBy. Callback wouldn't be triggered by nested create operations for relationship fields.

• #6236 1866286 Thanks @angrykoala! - Fix bug with some filters not properly working on subscriptions.

@neo4j/introspector@5.0.0

Major Changes

• #5837 721691a Thanks @MacondoExpress! - Changed how "@neo4j/introspector" generates list fields that now are generated as a list of non-nullable elements, as a list of nullable elements is not supported by Neo4j.

@neo4j/graphql@7.0.0

Major Changes

• #6153 0779691 Thanks @angrykoala! - Fails schema validation if a field with @relationship targets a type without @node.

  For example, the following schema will fail:

  type Movie @node {
      someActors: [Actor!]! @relationship(type: "ACTED_IN", direction: OUT)
  }
  
  type Actor {
      name: String
  }

• #5899 7335d8f Thanks @darrellwarde! - Nested mutation operations now follow the relationship direction behaviour as defined in queryDirection

• #6048 c667618 Thanks @darrellwarde! - Subscriptions are now an opt-in feature which can be enabled by using the @subscription directive on either schema or type.

  For example, to enable subscriptions for the whole schema (equivalent to before this breaking change):

  type Movie @node {
      title: String!
  }
  
  extend schema @subscription

  To enable subscriptions just for the Movie type:

  type Movie @node @subscription {
      title: String!
  }

• #5997 a716ef8 Thanks @angrykoala! - Remove publish method from Neo4jGraphQLSubscriptionsEngine interface as it is no longer used with CDC-based subscriptions. Implementing this method on custom engines will no longer have an effect, and it is no longer possible to call publish directly on Neo4jGraphQLSubscriptionsCDCEngine

• #5872 925ad8d Thanks @angrykoala! - Remove @private directive. This directive was intended to be used with the library @neo4j/graphql-ogm which is no longer supported.

• #5895 6afcadd Thanks @angrykoala! - Fails schema generation if there are conflicting plural names in types. For example, the following schema will fail, due to ambiguous Techs plural

  type Tech @node(plural: "Techs") {
      name: String
  }
  
  type Techs {
      value: String
  }

• #5927 530d8cd Thanks @MacondoExpress! - The typename_IN filter has been renamed to typename.

• #6077 4cf7c07 Thanks @darrellwarde! - Values specified within the @coalesce directive are now also returned when selecting those fields, and not just when those fields are used in a filter.

• #5755 9c75f92 Thanks @angrykoala! - Remove support for connectOrCreate operations

• #6158 19dc0dd Thanks @angrykoala! - Remove option CONNECT_OR_CREATE from argument nestedOperations from @relationship directives as it is no longer relevant after connectOrCreate have been removed

• #5898 a912404 Thanks @MacondoExpress! - Aggregations are no longer generated for ID fields.

• #5936 d48ea32 Thanks @mjfwebb! - Changes the result projection where there are multiple relationships between two nodes.

  In the case of using the connection API then multiple relationships will still be represented, as there is the ability to select the relationship properties. In the non-connection API case, the duplicate results will only return distinct results.

• #6109 6801d70 Thanks @darrellwarde! - The where field for nested update operations has been moved within the update input field.
  The where in its previous location was a no-op for all nested operations apart from update.

  For example, the following syntax would filter the Post nodes to update in Version 6:

  mutation {
      updateUsers(
          where: { name: { eq: "Darrell" } }
          update: {
              posts: {
                  where: { node: { title: { eq: "Version 7 Release Notes" } } }
                  update: { node: { title: { set: "Version 7 Release Announcement" } } }
              }
          }
      )
  }

  In Version 7, this where has been moved inside the update operation:

  mutation {
      updateUsers(
          where: { name: { eq: "Darrell" } }
          update: {
              posts: {
                  update: {
                      where: { node: { title: { eq: "Version 7 Release Notes" } } }
                      node: { title: { set: "Version 7 Release Announcement" } }
                  }
              }
          }
      )
  }

• #6124 d9cfd69 Thanks @angrykoala! - Fixed incorrect behavior of single and some filters on relationships to unions.

  Given the following union and relationship:

  union Production = Movie | Series

  and a relationship to this union:

  type Actor @node {
      name: String!
      actedIn: [Production!]! @relationship(type: "ACTED_IN", direction: OUT)
  }

  These queries previously returned incorrect results:

  query {
      actors(
          where: {
              actedIn: { single: { Movie: { title_CONTAINS: "The Office" }, Series: { title_ENDS_WITH: "Office" } } }
          }
      ) {
          name
      }
  }

  query {
      actors(
          where: {
              actedIn: { some: { Movie: { title_CONTAINS: "The Office" }, Series: { title_ENDS_WITH: "Office" } } }
          }
      ) {
          name
      }
  }

  Previously, conditions inside single and some were evaluated separately for each concrete type in the union, requiring all to match. This was incorrect.

  New behavior:

  • single: Now correctly returns actors with exactly one related node across the whole union, rather than per type.

  • some: Now correctly returns actors with at least one matching related node of any type in the union.

    This fix also applies to the deprecated filters actedIn_SINGLE and actedIn_SOME.

• #5931 5ce7d1d Thanks @darrellwarde! - DateTime and Time values are now converted from strings into temporal types in the generated Cypher instead of in server code using the driver. This could result in different values when the database is in a different timezone to the GraphQL server.

• #5778 56022ba Thanks @darrellwarde! - The deprecated directed argument has been removed, and queryDirection now only accepts two possible values - DIRECTED (default) and UNDIRECTED.

  Additionally, the directedArgument setting of excludeDeprecatedFields has been removed as these deprecated fields have been removed.

• #5976 7ddde75 Thanks @angrykoala! - Sets addVersionPrefix to true by default, this will prepend the Cypher version to all queries by default, ensuring that the correct Cypher version is used in Neo4j:

  CYPHER 5
  MATCH(this:Movie)

  This may be incompatible with older versions of Neo4j and can be disabled by setting cypherQueryOption.addVersionPrefix in the context to false:

  {
      cypherQueryOptions: {
          addVersionPrefix: true,
      },
  }

  For example, for an apollo server:

  await startStandaloneServer(server, {
      context: async ({ req }) => ({
          req,
          cypherQueryOptions: {
              addVersionPrefix: false,
          },
      }),
      listen: { port: 4000 },
  });

• [#5819](http...

@neo4j/graphql@6.6.1

Patch Changes

• #6171 c75cda3 Thanks @angrykoala! - Cypher optimized for connections only containing aggregate fields. The following query:

  {
      moviesConnection {
          aggregate {
              count {
                  nodes
              }
          }
      }
  }

  Will generate the new Cypher:

  CALL {
      MATCH (this:Movie)
      RETURN { nodes: count(DISTINCT this) } AS var0
  }
  RETURN { aggregate: { count: var0 } } AS this

  Instead of the less performant previous Cypher:

  CALL {
      MATCH (this:Movie)
      RETURN { nodes: count(DISTINCT this) } AS var0
  }
  CALL {
      WITH *
      MATCH (this1:Movie)
      WITH collect({ node: this1 }) AS edges
      WITH edges, size(edges) AS totalCount
      CALL {
          WITH edges
          UNWIND edges AS edge
          WITH edge.node AS this1
          RETURN collect({ node: { __id: id(this1), __resolveType: "Movie" } }) AS var2
      }
      RETURN var2, totalCount
  }
  RETURN { edges: var2, totalCount: totalCount, aggregate: { count: var0 } } AS this