I'm having an issue with sequelize on create with associated model with an "unique" value. It works when I create a "book" with a new "category", but if the "category" already exists it throws an error because it tried to duplicate an "unique" value. Is there a way to make sequelize.create try to use findAndCreate on a associated model?
My setup
Book:
const Book = sequelize.define('book', {
title: Sequelize.STRING,
})
Category:
const Category = sequelize.define('category', {
id: Sequelize.BIGINT,
name: {
type: Sequelize.STRING,
unique: true,
},
})
and the relationship:
Book.belongsTo(Category, { foreignKey: 'category_id' })
So, when i create a book i can write:
Book.create(
{
title: 'Book',
category: {
name: 'Foo'
}
}, {
include: [
{
model: Category
}
]
}
)
if after that I create another book:
Book.create(
{
title: 'Another Book',
category: {
name: 'Foo'
}
}, {
include: [
{
model: Category
}
]
}
)
it throws the error saying category.name must be unique
For anyone with the same question, this was my final solution:
On my services.js I created the function create that receives a request(req) and all my models already initialized.
create: (req, models) => {
const { category, ...rest } = req.body
if (category && category.name && !rest.category_id) {
return models.Category
.findOrCreate({
where: {
name: category.name
}
})
.then(([returnedCategory]) => {
return models.Book.create(
{
...rest,
category_id: returnedCategory.id
},
{ ...defaultOptions(models) }
)
})
} else {
return models.Book.create(rest, { ...defaultOptions(models) })
}
},
defaultOptions looks like this:
const defaultOptions = models => {
return {
attributes: {
exclude: ['category_id', 'created_at', 'updated_at'],
},
include: [
{ model: models.Category, attributes: ['id', 'name'] },
],
}
}
Related
I have end-point which is supposed to delete record from DB:
delete: async(roleId, actionId) => {
const actionrole = await ActionRoleModel.findAll({
where: {
roleId: roleId,
actionId: actionId
},
});
return await actionrole[0].destroy();
}
That [0] has to be here, because actionrole looks like [{...}].And here is the model:
'use strict';
module.exports = (sequelize, DataTypes) => {
var ActionRole = sequelize.define('actionroles', {
actionId: {
type: "UNIQUEIDENTIFIER",
field: "actionid"
},
roleId: {
type: "UNIQUEIDENTIFIER",
field: "roleid"
},
createdAt: {
field: "createdat",
type: DataTypes.DATE
},
updatedAt: {
field: "updatedat",
type: DataTypes.DATE
},
}, {});
ActionRole.associate = function(models) {
// associations can be defined here
};
ActionRole.removeAttribute('id');
return ActionRole;
};
But as an error in terminal I get
DatabaseError [SequelizeDatabaseError]: syntax error at or near "IN"
And here is SQL:
DELETE FROM "actionroles"
WHERE IN (
SELECT FROM "actionroles"
WHERE "roleid" = '53549d62-cd2a-497f-9d1c-1ee1901261ab' AND "actionid" = '6c70bf65-30fd-4640-91d0-8fbda85c4dd5'
LIMIT 1)
What's wrong? How can I fix that?
For anyone using Sequelize version 3 and above it looks like:
Model.destroy({
where: {
// conditions
}
})
So, in this case it would be look like this:
return await ActionRoleModel.destroy({
where: {
roleId: roleId,
actionId: actionId
}
});
And it works!
Imagine you have this data structure:
const data = {
posts: [{
id: 1,
title: "Post 1"
slug: "post-1"
}, {
id: 2,
title: "Post 2"
slug: "post-2"
}],
comments: [{
id: 1,
postId: "post-1",
text: "Comment 1 for Post 1"
}, {
id: 2,
postId: "post-1",
text: "Comment 2 for Post 1"
}, {
id: 3,
postId: "post-2",
text: "Comment 1 for Post 2"
}]
}
An you have the following route /posts/[postId[/[commentId]
so the Next.js structure folder is: posts/[postId]/[commented].js
Then you need to generate the static paths for this routes.
I'm coded the following:
export async function getStaticPaths() {
const { posts, comments } = data
const paths = posts.map((post) => {
return comments
.filter((comment) => comment.postId === post.slug)
.map((comment) => {
return {
params: {
postId: post.slug,
commentId: comment.id
}
}
})
})
}
But it's not working. The throwed error was:
Error: Additional keys were returned from `getStaticPaths` in page "/clases/[courseId]/[lessonId]". URL Parameters intended for this dynamic route must be nested under the `params` key, i.e.:
return { params: { postId: ..., commentId: ... } }
Keys that need to be moved: 0, 1.
How I can "map" or "loop" the data to a proper returned format?
Thanks in advance!
The problem seems to be that your returning this from getStaticPaths data with a wrong shape:
[
[ { params: {} }, { params: {} } ],
[ { params: {} } ]
]
The correct shape is:
[
{ params: {} },
{ params: {} },
{ params: {} }
]
Just tried this and it works.
export async function getStaticPaths() {
const paths = data.comments.map((comment) => {
return {
params: {
postId: comment.postId,
commentId: comment.id
}
}
});
console.log(paths);
return {
paths,
fallback: false
}
};
It generates 3 urls:
/posts/post-1/1
/posts/post-1/2
/posts/post-2/3
Is that what you need?
Like mention #Aaron the problem is for double array of filter y el map.
return {
paths: [
{ params: { id: '1' } },
{ params: { id: '2' } }
],
fallback: ...
}
Doc 📚 ➡ https://nextjs.org/docs/basic-features/data-fetching#the-paths-key-required
I have this promise chain in my code and it works just fine. Document actually has a value and is not null
...
.then(() => {
return db.document.findOne({
where: {
id: _document.get('id', req.transaction)
},
include: [{
model: db.documentChildren,
attributes: ['id', 'reference', 'uri', 'contentType', 'type', 'page']
},
{
model: db.tag,
attributes: ['id', 'key', 'value'], // We don't want meta columns
through: { attributes: [] } // Exclude join table
}],
transaction: req.transaction
})
})
.then(document => {
console.log('document = ', document)
...
Now I want to abstract that query into a function so that it can be reused.
I would have thought this would work, but for some reason document is always null and when I run the generated query it does have a result.
Why is document null when abstracting this query into its own function?
function findOneDocumentQuery (db, id, transaction) {
return db.document.findOne({
where: {
id: id
},
include: [{
model: db.documentChildren,
attributes: ['id', 'reference', 'uri', 'contentType', 'type', 'page']
},
{
model: db.tag,
attributes: ['id', 'key', 'value'], // We don't want meta columns
through: { attributes: [] } // Exclude join table
}],
transaction: transaction
})
}
...
.then(() => {
return findOneDocumentQuery(db, _document.get('id', req.transaction))
})
.then(document => {
console.log('document = ', document)
...
I think the bracket to findOneDocumentQuery is wrong, it should be:
return findOneDocumentQuery(db, _document.get('id'), req.transaction);
Trying to query by all association's attribute, but get all associations
# FAQs: { id: 1, name: 'How to do it?' }, { id: 2, name: 'How to FIX it?' }
# tags: { id: 1, slug: 'api' }, { id: 2, slug: 'beta' }
# taggings: { id: 1, faqId: 1, mainEntityId: 1, mainEntityType: 'faq' }, { id: 2, faqId: 1, mainEntityId: 2, mainEntityType: 'faq' }
const query = { slugs: ['api'] }
const foundFAQs = await this.models.FAQ.findAll({
where: {
'$taggings.tag.slug$': { $in: query.slugs },
},
include: [{
model: this.models.Tagging,
as: "taggings",
include: [{
model: this.models.Tag,
as: 'tag',
}],
}],
})
My model definition:
models.Tagging.belongsTo(models.Tag, { as: 'tag', onDelete: 'cascade' });
models.Tag.hasMany(models.Tagging, { as: 'taggings', onDelete: 'cascade' });
models.Tag.belongsToMany(models.FAQ, { through: models.Tagging, as: 'faqs' });
models.FAQ.hasMany(models.Tagging, { as: 'taggings', onDelete: 'cascade' });
models.FAQ.belongsToMany(models.Tag, { through: models.Tagging, as: 'tags' });
What do you expect to happen?
I want to get all FAQs what has associated TAG api and have all its tags.
Object:
{
id: 1,
name: 'How to do it?',
tags: [
{ id: 1, slug: 'api' },
{ id: 2, slug: 'beta' }
]
}
What is actually happening?
Query returns object:
{
id: 1,
name: 'How to do it?',
tags: [
{ id: 1, slug: 'api' }
]
}
Output
SELECT \"faq\".\"id\", \"faq\".\"name\", \"faq\".\"bankId\", \"faq\".\"priority\",
\"faq\".\"publishedLocales\", \"faq\".\"createdAt\", \"faq\".\"updatedAt\", \"taggings\".\"id\"
AS \"taggings.id\", \"taggings\".\"tagId\" AS \"taggings.tagId\", \"taggings\".\"locked\"
AS \"taggings.locked\", \"taggings\".\"sdkId\" AS \"taggings.sdkId\", \"taggings\".\"guideId\"
AS \"taggings.guideId\", \"taggings\".\"newsId\" AS \"taggings.newsId\", \"taggings\".\"faqId\"
AS \"taggings.faqId\", \"taggings\".\"apiId\" AS \"taggings.apiId\", \"taggings\".\"createdAt\"
AS \"taggings.createdAt\", \"taggings\".\"updatedAt\" AS \"taggings.updatedAt\", \"taggings->tag\".\"id\"
AS \"taggings.tag.id\", \"taggings->tag\".\"name\" AS \"taggings.tag.name\", \"taggings->tag\".\"slug\"
AS \"taggings.tag.slug\", \"taggings->tag\".\"tagType\" AS \"taggings.tag.tagType\", \"taggings->tag\".\"mainEntityId\"
AS \"taggings.tag.mainEntityId\", \"taggings->tag\".\"createdAt\"
AS \"taggings.tag.createdAt\", \"taggings->tag\".\"updatedAt\"
AS \"taggings.tag.updatedAt\" FROM \"faqs\" AS \"faq\" INNER JOIN \"taggings\"
AS \"taggings\" ON \"faq\".\"id\" = \"taggings\".\"faqId\" LEFT OUTER JOIN \"tags\"
AS \"taggings->tag\" ON \"taggings\".\"tagId\" = \"taggings->tag\".\"id\" WHERE \"faq\".\"bankId\" = 'bank.csas'
AND \"taggings->tag\".\"slug\" IN ('faq') ORDER BY \"faq\".\"priority\" DESC;
Dialect: postgres
Dialect version: pg#^6.1.0
Database version: PostgreSQL 10.1
Sequelize version: 4.23.2
Tested with the latest release: No (4.23.2)
Node: 8.6.0
I am not sure if it is a bug or I'm doing something wrong.
Thanks
You're looking to include all related data for each record, but filter those records on that related data. You're going to want to get Sequelize to generate SQL similar to:
SELECT "faq"."id", ....
FROM "faqs" AS "faq"
INNER JOIN "taggings" AS "taggings" ON "faq"."id" = "taggings"."faqId"
LEFT OUTER JOIN "tags" AS "taggings->tag" ON "taggings"."tagId" = "taggings->tag"."id"
LEFT OUTER JOIN "tags" AS "taggings->tagdata" ON "taggings"."tagId" = "taggings->tagdata"."id"
WHERE "faq"."bankId" = 'bank.csas' AND "taggings->tag"."slug" IN ('faq')
ORDER BY "faq"."priority" DESC;
Something like this might work:
const foundFAQs = await this.models.FAQ.findAll({
where: {
'$taggings.tag.slug$': { $in: query.slugs },
},
include: [{
model: this.models.Tagging,
as: "taggings",
include: [{
model: this.models.Tag,
as: 'tag',
},{
model: this.models.Tag,
as: 'tagdata',
}],
}],
})
The idea here is that you're performing a join on which you are filtering, and then another join to get the extra records for the filtered rows.
This worked for me today -- nested includes:
getReferralPlanForThisCode(parent, args, context) {
let {referralCode} = args;
return Promise.resolve()
.then(() => {
let referralPlan = connectors.ReferralPlans.findAll({
include: [{
model: connectors.ReferralCodes,
where: {unique_referral_code: referralCode},
include: [{
model: connectors.epUserData, as: 'referrer',
}],
}],
})
return referralPlan;
})
.then(referralPlan => {
return referralPlan;
})
.catch((err) => {
console.log(err);
});
}
I have been writing an API that uses GraphQL. I am still pretty new to it, and have been running into some problems regarding mutations. A simplistic form of my API has two record types. There is a contact record and a tag record. A contact record can have multiple tag records associated with it.
The schema I wrote for each of these record types are below:
const Tag = new graphQL.GraphQLObjectType({
name: 'Tag',
description: 'Categorizes records into meaningful groups',
fields: () => ({
_id: {
type: graphQL.GraphQLID
},
name: {
type: graphQL.GraphQLString
}
})
});
const Contact = new graphQL.GraphQLObjectType({
name: 'Contact',
description: 'Contact record',
fields: () => ({
_id: {
type: graphQL.GraphQLID
},
name: {
type: graphQL.GraphQLString
},
tags: {
type: new graphQL.GraphQLList(Tag),
resolve: function(src, args, context) {
return TagModel.findByContactId(src._id)
.then(tags => {
return Promise.map(tags, (tag) => {
return TagModel.findById(tag.tag_id);
});
});
}
}
})
});
I can make a mutation easy enough on records such as tags since they don't contain nested records of their own, but I'm not sure how to make a mutation on a record like contacts since it can contain tags as well. The mutation code I put in place looks like this:
const Mutation = new graphQL.GraphQLObjectType({
name: 'Mutation',
fields: {
createContact: {
type: Contact,
description: "Create Contact",
args: {
name: {type: new graphQL.GraphQLNonNull(graphQL.GraphQLString)},
tags: {type: new graphQL.GraphQLList(Tag)}
},
resolve: function(source, args) {
return ContactModel.save(args.name);
}
}
}
});
I'm not sure how to complete the resolver in the mutation in order to be able to save a contact and tag records at the same time. For instance, if I made a mutation query to save a new contact record with a new tag like this:
{"query": "mutation createNewContact {
contact: createContact (name: "John Smith", tags { name: "family" } )
{_id, text, tags { name } } }" }
Is there something special that I need to do in my mutation schema in order to allow for this type of mutation to happen?
You can't use Tag as an input object type, you would have to create a type like TagInput
const TagInput = new GraphQLInputObjectType({
name: 'TagInput',
fields: {
_id: { type: GraphQLID },
name: { type: GraphQLString }
}
});
It is recommended to always create Input version of your normal type. You could do the same with Contact by creating ContactInput. Then you could create a mutation in very similar way you did it
const Mutation = new GraphQLObjectType({
name: 'Mutation',
fields: {
createContact: {
type: Contact,
args: {
contact: { type: new GraphQLNonNull(ContactInput) },
tags: { type: new GraphQLList(TagInput) }
},
resolve: (root, args, context) => {
console.log(args);
// this would console something like
// { contact: { name: 'contact name' },
// tags: [ { name: 'tag#1' }, { name: 'tag#2' } ] }
// here create contact with tags
}
}
});
The query you would run would look like that
{
"operationName": "createContact",
"query": "mutation createContact($contact: ContactInput!, $tags: [TagInput])
{
createContact(contact: $contact, tags: $tags) {
_id
text
tags {
name
}
}
}",
"variables": {
contact: { name: "contact name" },
tags: [ { name: "tag#1" }, { name: "tag#2" } ]
}
}