Version: Next

API Reference

The migrateDB method takes the three arguments, listed below.

  • config: Knex database configuration options.
  • schemaText: GraphQL schema text.
  • options:
    • dbSchemaName (default: 'public'): table schema: <schemaName>.<tableName>.
    • dbTablePrefix (default: ''): table name prefix: <prefix><tableName>.
    • dbColumnPrefix (default: ''): column name prefix: <prefix><columnName>.
    • updateComments (default: false): by default, migrateDB will not create or update comments on table and columns.
    • scalarMap (default: undefined): Custom Scalar mapping
    • mapListToJson (default: true): Map scalar/enum lists to json column type by default.
    • plugins (default: []): List of graphql-migrations plugins which describes queries that can be executed during migrations.
    • debug (default: false): displays debugging informations and SQL queries.
    • removeDirectivesFromSchema: (default: true): Strips all directives from schema before processing.
    • operationFilter: Filter out database operations that we don't want.

Table and Column Comments#

When updateComments is set to true it will automatically create/update table and column comments. Table / column comments are type or field description parsed from the model schema. The parsing logic takes care of stripping off all annotations leaving only business related comments.

The below model:

Note business model table comment.
type Note {
id: ID!
Note title.
title: String!

Execution with the updateComments options set to true, as shown below

migrateDB(dbConfig, schemaText, {
updateComments: true
}).then((ops) => {

Will create / update the note table comments as shown below:

List of relations
Schema | Name | Type | Owner | Size | Description
public | note | table | postgresql | 8192 bytes | Note business model table comment.
public | note_id_seq | sequence | postgresql | 8192 bytes |
(2 rows)
\d+ note
Table "public.note"
Column | Type | Modifiers | Storage | Stats target | Description
title | character varying(255) | not null | extended | | Note title.
id | integer | not null default nextval('note_id_seq'::regclass) | plain | |
"note_pkey" PRIMARY KEY, btree (id)

NOTE: This options may not be supported by some database.

Scalar Mapping#

This option is a function used to specify custom Scalar mapping. It takes three arguments (field: GraphQLField, fieldType: GraphQLScalarType, dbAnnotation: any), depending on the field type you could return the TableColumnTypeDescriptor:

interface TableColumnTypeDescriptor {
* Knex column builder function name.
type: TableColumnType | string
* Builder function arguments.
args: any[]

For example, if you want a custom CustomDate scalar to be have a date type in the database:

import { migrateDB } from 'graphql-migrations';
const schemaText = ```
""" @model """
type Note {
id: ID!
title: String!
createdAt: CustomDate
scalar CustomDate
const dbConfig = {
// Knex.js db configuration
migrateDB(dbConfig, schemaText, {
scalarMap: (
field: GraphQLField<any, any>,
scalarType: GraphQLScalarType,
annotations: any,
): TableColumnTypeDescriptor | null => {
// custom date converted to date type
if ( === 'CustomDate') {
return {
type: 'date',
args: []
return undefined
}).then((ops) => {

A good example of this function is the default scalar type to database column function, that Graphback uses to convert scalar types based on certain creteria.


Checkout the Graphback Scalars chapter to learn more of supported GraphQL scalars with automatic mapping to the underlying database type.