currently i'm creating a GraphQL implementation of an existing API.
i want to create Search Function not with primary column, which is not id, but id_position.
here is my scheme :
type Query #guard {
subordinate(
positionId: ID
): [Employee!]! #all
}
type Position #key(fields: "id") {
id: ID
name: String
}
"Account of a person who utilizes this application."
type Employee #key(fields: "id") {
id: ID
name: String
id_position: Int
}
but, when I run this :
query EmployeeSubordinate($id: ID) {
subordinate(positionId: $id) {
name
}
}
{
"id" : 93
}
I ve got result all rows of employee, not employee with id_position = 93
how to solve this?
I think the problem is here
type Query #guard {
subordinate(
positionId: ID
): [Employee!]! #all
}
The #all is getting all records from DB #all docs
The correct way is this
type Query #guard {
subordinate(
positionId: ID
): Employee
}
I removed ! from Employee because your ID is not required, so in some cases it masy return null, i you handle like thatin backend.
And alsoe I removed [] because you no longer getting many Employee you just getting one.
Related
Maybe the title is not accurate but I really don't know how to describe it anymore. I went through multiple documentations and descriptions but still couldn't figure it out.
I want to implement a basic social media like followers/following query on my type User. I am using MySQL and for that I made a separate table called Follow as it's a many-to-many connection.
Here is a pseudo-ish representation of my tables in the database without the unnecessary columns:
Table - User
user_id primary key Int
Table - Follow
follow_er foreign_key -> User(user_id) Int
follow_ed foreign_key -> User(user_id) Int
A user could "act" as a follow_er so I can get the followed people
And a user could be follow_ed, so I can get the followers.
My prisma schema look like this:
model User {
user_id Int #id #default(autoincrement())
following Follow[] #relation("follower")
followed Follow[] #relation("followed")
}
model Follow {
follow_er Int
follower User #relation("follower", fields: [follow_er], references: [user_id])
follow_ed Int
followed User #relation("followed", fields: [follow_ed], references: [user_id])
##id([follow_er, follow_ed])
##map("follow")
}
By implementing this I can get the followers and following object attached to the root query of the user:
const resolvers = {
Query: {
user: async (parent, arg, ctx) => {
const data = await ctx.user.findUnique({
where: {
user_id: arg.id
},
include: {
following: true,
followed:true
}
})
return data
}....
Here is my GraphQL schema I tried to make:
type Query{
user(id: Int!): User
}
type User{
id: ID
following: [User]
followed: [User]
}
So I can get something like:
query {
user(id: $id) {
id
following {
id
}
followed{
id
}
}
}
}
But I couldn't make it work as even if I get the the array of objects of {follow-connections}:
[
{
follow_er:1,
follow_ed:2
},
{
follow_er:1,
follow_ed:3
},
{
follow_er:3,
follow_ed:1
},
]
I can't iterate through the array. As far as I know, I have to pass either the follow_er or follow_ed, which is a user_id to get a User object.
What am I missing? Maybe I try to solve it from a wrong direction. If anybody could help me with this, or just tell me some keywords or concepts I have to look for it would be cool. Thanks!
I would suggest creating self-relations for this structure in the following format:
model User {
id Int #id #default(autoincrement())
name String?
followedBy User[] #relation("UserFollows", references: [id])
following User[] #relation("UserFollows", references: [id])
}
And then querying as follows:
await prisma.user.findUnique({
where: { id: 1 },
include: { followedBy: true, following: true },
})
So you will get a response like this:
I get this error while querying:
"Cannot return null for non-nullable field Transaction.createdAt."
This is the query:
query getMyTransactions {
getMyTransactions {
id
createdAt
}
}
The schema for this query is:
extend type Transaction #key(fields: "id") {
id: ID! #external
}
type Query {
getMyTransactions: [Transaction!]!
}
And the other schema has Transaction type:
type Transaction #key(fields: "id") {
id: ID!
createdAt: String!
}
What's the problem?
EDIT: If I query for:
getMyTransactions {
id
}
Works ok I and get all the id of the query, but if I include another attribute the query fails.
In simple words - you declare createdAt type to be non-null value, but somewhere in your data createdAt is null.
createdAt: String!
! after the type name. This means that our server always expects to
return a non-null value for this field, and if it ends up getting a
null value that will actually trigger a GraphQL execution error,
letting the client know that something has gone wrong. GraphQl docs
Example
For this remote/local data (createdAt missing from the first object):
const Transaction = [
{
id: "1",
/* createdAt: "01/09/2020" */ /* missing from the data */
},
{
id: "2",
createdAt: "02/09/2020"
}
];
Executing query
query{
getMyTransactions {
id
createdAt
}
}
throw error:
"message": "Cannot return null for non-nullable field
Transaction.createdAt.",
To solve this:
Option 1: Add some resolver logic and/or add validation related to your required data.
Option 2: Remove the require from createdAt (return null allowed):
I have a User table that contains user data. I also have a Relationship table that contains a parent_id and a child_id as well as the incrementing id. Each Parent can have many Children and I am trying to retrieve the array of children using the parent_id. I am currently getting an array, but the array only contains the Relationship table data.
How do I get the data from the User table? I was under the impression that because there is a relationship the following should work, but it doesnt.
query {
getChildren {
user{
id
name
email
}
}
}
The relevant code is as follows:
Query:
export const getChildren = queryField('getChildren', {
type: 'User',
list: true,
resolve: async (_parent, {}, ctx) => {
const userId = getUserId(ctx);
if (!userId) {
// TODO -think I might need to throw an error here
return;
}
const children = await ctx.prisma.relationship.findMany({
where: {
parent_id: userId,
},
});
return children;
},
});
schema.prisma:
model Relationship {
child_id Int
id Int #default(autoincrement()) #id
parent_id Int
child User #relation("Relationship_child_idToUser", fields: [child_id], references: [id])
parent User #relation("Relationship_parent_idToUser", fields: [parent_id], references: [id])
}
model User {
created_at DateTime #default(now())
email String? #unique
id Int #default(autoincrement()) #id
ischild Boolean #default(false)
name String?
password String
children Relationship[] #relation("Relationship_child_idToUser")
parents Relationship[] #relation("Relationship_parent_idToUser")
}
The query below should do it
prisma.relationship.findMany({
where: {
parent_id: user_id,
},
include: {
child: true,
},
})
You need to specify the include param explicitly to include the relationship in the output.
I need help figuring out the GraphQL tag for use with Apollo Client. The Docs don't go far beyond the most basic use case for mutations.
My goal is to have the only required input be an email. If the other variables are present, I would like those to be accepted and create a proposal with all that information.
I have the mutation (in both only email and full variables scenarios) working successfully on the GraphQl Playground (if it helps, you can find it here and test it out, look at the schema, etc.,): https://prisma2-graphql-yoga-shield.now.sh/playground)
mutation {
createOneProposal(
data: {
email: "fake#gmail.com"
name: "Sean"
types: {
create: {
model: PURCHASE
name: "e-commerce"
cost: 600
services: {
create: [
{ service: "Responsive web design" }
{ service: "Another service!" }
{ service: "And yet another service!" }
]
}
}
}
}
) {
created_at
proposal_id
types {
cost
model
name
type_id
services {
service
service_id
}
}
}
}
Producing as a result:
{
"data": {
"createOneProposal": {
"created_at": "2020-02-27T21:28:53.256Z",
"proposal_id": 35,
"types": [
{
"cost": 600,
"model": "PURCHASE",
"name": "e-commerce",
"type_id": 6,
"services": [
{
"service": "Responsive web design",
"service_id": 10
},
{
"service": "Another service!",
"service_id": 11
},
{
"service": "And yet another service!",
"service_id": 12
}
]
}
]
}
}
}
My initial design for the gql tag:
export const NEW_PROPOSAL = gql`
mutation createOneProposal(
$email: String!
$name: String
$cost: Int
$model: Model
$service: Service
) {
createOneProposal(
email: $email
name: $name
cost: $cost
model: $model
service: $service
) {
created_at
proposal_id
types {
services {
service_id
}
}
}
}
`;
But, I get a lot of errors with this.
{"errors":[
{"Variable "$service" cannot be non-input type `"Service`".","locations":[{"line":1,"column":97}]},
{"Unknown argument "email" on field "createOneProposal`" of type "Mutation`".","locations":[{"line":2,"column":21}]},
{"Unknown argument "name" on field "createOneProposal`" of type "Mutation`".","locations":[{"line":2,"column":36}]},
{"Unknown argument"cost" on field "createOneProposal\" of type "Mutation`".","locations":[{"line":2,"column":49}]},
{"Unknown argument "model" on field "createOneProposal`" of type "Mutation`".","locations":[{"line":2,"column":62}]},
{"Unknown argument "service" on field "createOneProposal`" of type "Mutation`".","locations":[{"line":2,"column":77}]},
{"Field "createOneProposal" argument "data" of type "ProposalCreateInput!`" is required, but it was not provided.","locations":[{"line":2,"column":3}]}]}
So, how can I go about this... I figured out the query version (much easier...), but I just can't figure this out!
My schema, if it helps:
generator client {
provider = "prisma-client-js"
}
datasource db {
provider = "mysql"
url = env("MYSQL_URL_PRISMA2")
}
model Post {
content String #default("")
created_at DateTime #default(now())
post_id Int #default(autoincrement()) #id
published Boolean #default(false)
published_at DateTime?
title String #default("")
author User
}
model Profile {
bio String?
profile_id Int #default(autoincrement()) #id
user_id User
}
model Proposal {
email String #unique
name String?
proposal_id Int #default(autoincrement()) #id
created_at DateTime #default(now())
types Type[]
}
model Type {
cost Int?
name String?
model Model? #default(SUBSCRIPTION)
services Service[]
type_id Int #default(autoincrement()) #id
proposal_id Proposal
}
model Service {
service_id Int #default(autoincrement()) #id
service String?
type_id Type
}
model User {
email String #default("") #unique
name String #default("")
password String #default("")
role Role #default(USER)
user_id Int #default(autoincrement()) #id
posts Post[]
profiles Profile[]
}
enum Role {
USER ADMIN
}
enum Model {
SUBSCRIPTION PURCHASE CUSTOM
}
GraphQL types are categorized as either input types or output types. Input types are used for inputs like variable definitions or argument definitions. Output types are used for typing fields, which are what compose the actual response. Certain types, like scalars and enums, can be used as either an input or an output. However, with objects, there are output object types (sometimes referred to just object types or objects) and input object types.
Service is an output type, so it can't be used where an input type is expected (in this case, a variable definition). Examine the schema generated by Prisma to determine the appropriate type to use.
Thanks to some very needed direction from #xadm, I figured out the structure of the tag! For anyone who comes across this in the future:
mutation createOneProposal($input: ProposalCreateInput!){
createOneProposal(data:$input){
created_at
name
email
proposal_id
type{
cost
description
model
name
type_id
services{
service
cost
service_id
}
}
}
}
I have a CHAT_MESSAGE_FRAGMENT that returns all the message data from my Hasura graphql api.
However, the Gifted Chat react-native component requires the data in a specific structure so I'm attempting to convert it with the query below.
I'm able to alias all the top level data but can't figure out how to add a nested level of data.
I'm guessing it isn't possible but I thought I'd ask in case I'm missing something.
const GIFTED_CHAT_GROUP_MESSAGES_QUERY = gql`
query chatGroupMessages($chatGroupId: Int!) {
chat_message(
where: { to: { id: { _eq: $chatGroupId } } }
) {
_id: id,
# user: {
# _id: from.id, <== How do I add
# name: from.name, <== this secondary level?
# },
text: message,
image: image_url,
createdAt: created_at,
system: message_type,
}
}
${CHAT_MESSAGE_FRAGMENT}
`;
Assuming you already have chat_message.user_id -> users.id foreign key constraint set up, you'll also need to alias the from object in addition aliasing any of its nested fields:
const GIFTED_CHAT_GROUP_MESSAGES_QUERY = gql`
query chatGroupMessages($chatGroupId: Int!) {
chat_message(
where: { to: { id: { _eq: $chatGroupId } } }
) {
_id: id,
from: user: {
_id: id,
name
},
text: message,
image: image_url,
createdAt: created_at,
system: message_type,
}
}
${CHAT_MESSAGE_FRAGMENT}
`;
The secondary level of data is basically nested object queries in Hasura. You can nest any number of queries as long as a relationship has been created.
In this case, assuming the chat_message table has a user_id field, you can establish a foreign key constraint for chat_message.user_id -> users.id, where users is a table with id as primary key.
Once the foreign key constraint is created, Hasura Console automatically suggests relationships. Here user would be an object relationship in chat_message table.
Here's the official docs link for Creating a relationship