post resolver fetches posts from postgres database and return them in response to graphql request.
The return value is array of Post entity objects. Post entity structure is given below.
{
id: 29,
title: 'Turkey Earthquake kills 1000 plus people',
text: "Body is saying and don't go to the dark path",
points: 0,
creatorId: 44,
createdAt: 2023-02-07T06:53:39.453Z,
updatedAt: 2023-02-08T07:01:54.409Z,
creator: User {
id: 44,
username: 'prashant',
email: 'prashant#gmail.com',
password: '$argon2id$v=19$m=65536,t=3,p=4$LqsF/Gb3E8J4+vm5TszyWQ$d+LhZ6mNOE4eLqWvdwItQ3qYMZr17+CfvPt9l0Iqk3M',
createdAt: 2023-02-06T07:32:11.717Z,
updatedAt: 2023-02-06T07:32:11.717Z
}
}
Now I have a subquery in SELECT section of the query.
#Query(() => PaginatedPosts)
async posts(
#Arg('limit', () => Int) limit: number,
#Arg('cursor', () => String, { nullable: true }) cursor: string | null,
#Ctx() { req, dataSource }: MyContext
): Promise<PaginatedPosts> {
const realLimit = Math.min(30, limit)
const realLimitPlusOne = realLimit + 1
const qb = dataSource
.createQueryBuilder()
.select('p')
.from(Post, 'p')
.innerJoinAndSelect('p.creator', 'u')
.orderBy('p.createdAt', 'DESC')
.take(realLimitPlusOne)
if (req.session.userId)
qb.addSelect((qb) => {
return qb
.subQuery()
.select('d.value')
.from(Updoot, 'd')
.where('"userId" = :userId AND "postId" = p.id', {
userId: req.session.userId
})
}, 'voteStatus')
else qb.addSelect('null', 'voteStatus')
if (cursor)
qb.where('p.createdAt < :cursor', {
cursor: new Date(parseInt(cursor))
})
const posts = await qb.getMany() // this is returning the above structure without `voteStatus` field in it.
console.log('posts: ', posts[0])
return {
posts: posts.slice(0, realLimit),
hasMore: posts.length === realLimitPlusOne
}
}
But instead of using await qb.getMany() if I use
const posts = await qb.getRawMany()
I get the data in following structure
{
p_id: 29,
p_title: 'Turkey Earthquake kills 1000 plus people',
p_text: "Body is saying and don't go to the dark path",
p_points: 0,
p_creatorId: 44,
p_createdAt: 2023-02-07T06:53:39.453Z,
p_updatedAt: 2023-02-08T07:01:54.409Z,
u_id: 44,
u_username: 'prashant',
u_email: 'prashant#gmail.com',
u_password: '$argon2id$v=19$m=65536,t=3,p=4$LqsF/Gb3E8J4+vm5TszyWQ$d+LhZ6mNOE4eLqWvdwItQ3qYMZr17+CfvPt9l0Iqk3M',
u_createdAt: 2023-02-06T07:32:11.717Z,
u_updatedAt: 2023-02-06T07:32:11.717Z,
voteStatus: -1
}
Which is not compatible with the graphql type I intend to respond with. So I am confused whether there is another way I can use queryBuilder to return the result with the compatible type or do I have to explicitly create a mapping function to map each object to desired structure of Post entity.
Related
I have a posts resolver which uses this logic to return posts:
const qb = Post.createQueryBuilder('post')
.select(['post.id AS id', 'post.createdAt AS id'])
.addSelect(
"json_build_object('id', user.id, 'username', user.username, 'email', user.email, 'createdAt', user.createdAt, 'updatedAt', user.updatedAt)",
'creator'
)
if (req.session.userId)
qb.addSelect((qb) => {
return qb
.subQuery()
.select('updoot.value')
.from(Updoot, 'updoot')
.where('"userId" = :userId AND "postId" = post.id', {
userId: req.session.userId
})
}, 'voteStatus')
else qb.addSelect('null', 'voteStatus')
qb.innerJoin('post.creator', 'user')
.orderBy('post.createdAt', 'DESC')
.take(realLimitPlusOne)
if (cursor)
qb.where('post.createdAt < :cursor', {
cursor: new Date(parseInt(cursor))
})
qb.printSql()
const posts = await qb.getRawAndEntities()
console.log('posts: ', posts['raw'][0])
return {
posts: posts['raw'].slice(0, realLimit),
hasMore: posts['raw'].length === realLimitPlusOne
}
I want my data to be structured in this way:
posts: {
id: 29,
createdAt: 2023-02-07T06:53:39.453Z,
creator: {
id: 44,
username: 'prashant',
email: 'prashant#gmail.com',
createdAt: '2023-02-06T13:02:11.717504',
updatedAt: '2023-02-06T13:02:11.717504'
},
voteStatus: -1
}
But instead of id and createdAt I get post_id and post_createdAt key names which is not compatible with my graphql types. Even though I added the aliases for these columns I am getting:
posts: {
post_id: 29,
post_createdAt: 2023-02-07T06:53:39.453Z,
creator: {
id: 44,
username: 'prashant',
email: 'prashant#gmail.com',
createdAt: '2023-02-06T13:02:11.717504',
updatedAt: '2023-02-06T13:02:11.717504'
},
voteStatus: -1
}
Is there any problem with my code or should I create a mapping function which key names to appropriate ones?
I'm trying to store history of workout in realm, my addHistory function looks like this
export function addHistory(workout, exercise, sets, _id) {
console.log({
workout,
exercise,
sets,
_id,
});
if (
_id !== undefined &&
workout !== undefined &&
exercise !== undefined &&
sets !== undefined
) {
// return console.log("HISTORY ", { workout, exercise, sets, _id });
return realm.write(() => {
return realm.create("workoutData", {
_id: _id,
exercise,
workout,
sets,
workoutDate: new Date(Date.now()),
});
});
} else {
alert("History is incomplete");
}
}
Schema of the workoutData is as follows:
exports.workoutData = {
name: "workoutData",
primaryKey: "_id",
properties: {
_id: "int",
workout: "workouts",
exercise: "exercise",
workoutDate: "date",
sets: "sets[]",
},
};
Now when I add sets and click on finishWorkoutHandler the logic works fine before the addHistory function but when addHistory is executed it throws the error as stated in the question.
//finish workout handler
const finishWorkoutHandler = () => {
if (sets.length == 0) {
return;
}
let setsFromRealm = realm.objects("sets");
let workoutData = realm.objects("workoutData");
let setsArray = [];
exercises.forEach((exercise) => {
sets
.filter((items) => items.exercise._id == exercise._id)
.forEach((sets) => {
let _id = 0;
if (setsFromRealm.length > 0) {
_id = realm.objects("sets").max("_id") + 1;
}
addSet(
sets.name,
parseInt(sets.weight),
parseInt(sets.reps),
parseInt(sets.rmValue),
sets.isHeighest,
sets.exercise,
_id,
sets.profile,
sets.failedSet,
sets.warmupSet,
sets.notes
);
let indiSet = {
name: sets.name,
weight: parseInt(sets.weight),
reps: parseInt(sets.reps),
rmValue: parseInt(sets.rmValue),
isHeighest: sets.isHeighest,
_id: _id,
profile: sets.profile,
failedSet: sets.failedSet,
warmupSet: sets.warmupSet,
notes: sets.notes,
createdDate: new Date(Date.now()),
};
setsArray.push(indiSet);
});
let workoutDataId = 0;
let setsArrcopy = setsArray;
console.log("SETS ", realm.objects("sets"));
console.log("SETS ", setsArrcopy);
if (workoutData.length > 0) {
workoutDataId = realm.objects("workoutData").max("_id") + 1;
}
**WORKING AS EXPECTED TILL HERE**
// problem lies here
addHistory(params.workout, exercise, setsArrcopy, workoutDataId);
});
dispatch(setsEx([]));
goBack();
};
the structure of setsArrCopy containing sets is as follows
[
({
_id: 6,
createdDate: 2022-09-29T16:27:06.128Z,
failedSet: false,
isHeighest: false,
name: "Thai",
notes: "",
profile: [Object],
reps: 12,
rmValue: 64,
warmupSet: false,
weight: 56,
},
{
_id: 7,
createdDate: 2022-09-29T16:27:06.151Z,
failedSet: false,
isHeighest: false,
name: "Thsi 3",
notes: "",
profile: [Object],
reps: 10,
rmValue: 75,
warmupSet: false,
weight: 66,
})
];
the logic is also working fine in terms of assigning new ids to the sets being added in a loop. But somehow its throwing error when passing setArrCopy to addHistory function. Although its an array of sets not a single object?
I have a query:
createNotification: async (_, args, {req, res}) => {
const followedBy = await prisma.user.updateMany({
where: {
following: {
some: {
id: req.userId
},
},
},
data: {
notifications: {
create: {
message: args.message,
watched: false,
},
},
},
})
And User and Notification models:
model User {
id Int #id #default(autoincrement())
email String #unique
name String
user_name String #unique
password String
movies Movie[]
notifications Notification[]
followedBy User[] #relation("UserFollows", references: [id])
following User[] #relation("UserFollows", references: [id])
}
model Notification {
id Int #id #default(autoincrement())
link String?
movie_id Int?
message String
icon String?
thumbnail String?
user User #relation(fields: [userId], references: [id])
userId Int
watched Boolean
}
When I run my query I get an error:
Unknown arg `notifications` in data.notifications for type UserUpdateManyMutationInput. Did you mean `email`? Available args:
type UserUpdateManyMutationInput {
email?: String | StringFieldUpdateOperationsInput
name?: String | StringFieldUpdateOperationsInput
user_name?: String | StringFieldUpdateOperationsInput
password?: String | StringFieldUpdateOperationsInput
}
The strange thing is that this works:
const followedBy = await prisma.user.findUnique({
where: {id: req.userId},
include: {
followedBy: true,
},
});
followedBy.followedBy.map(async(user) => {
await prisma.user.update({
where: {id: user.id},
data: {
notifications: {
create: {
message: args.message,
watched: false,
},
},
},
});
});
But this isn't making the best of what Prisma offers.
As of September 2021, Prisma does not support mutating nested relations in a top-level updateMany query. This is what the typescript error is trying to tell you, that you can only access email, name, user_name and password fields inside data. There's an open feature request for this which you could follow if you're interested.
For the schema that you have provided, here's a possible workaround that's slightly less readable but more optimized than your current solution.
createNotification: async (_, args, {req, res}) => {
// get the followers array of req.userId
const followedBy = await prisma.user.findUnique({
where: { id: req.userId },
include: {
followedBy: true,
},
});
// array of notification objects to be created, one for each follower of req.userId
let messageDataArray = followedBy.followedBy.map((user) => {
return {
userId: user.id,
message: args.message,
watched: false,
};
});
// do a bulk createMany.
// Since it's one query, it should be more optimized than running an update for each user in a loop.
await prisma.notification.createMany({
data: messageDataArray,
});
};
If you're interested, here's the docs reference for the kinds of nested updates that are possible.
How do I update my database using Postman?
I'm trying to learn Express using Postman to test different types of requests, but I can't figure out how to update my data set. This is the structure of the data set I'm using.
var DB = [
{ category: 'Pets',
products: [
{ name: 'banjo',
color: 'grey',
mean: true,
description: "meows"
},
{ name: 'rigby',
color: 'black and white',
mean: false,
description: 'barks'
}]}]
Lets say I want to add another pet into the pets category so in products
{name: frank, color: orange, : mean: false: description: glubs}
I cant figure out how to add it correctly in Postman so that itll update. My code for the update is as follow:
app.post("/product/add", (req, res) => {
var category = req.body.category;
const { name, color, mean, description } = req.body;
var product = { name:name, color:color, mean:mean, description:description }; console.log(product);
var index = DB.findIndex(x => x.category == category);
if(index !== -1){
var indexProduct = DB[index].products.findIndex(x => x.name == product.name); if(indexProduct !== -1){
DB[index].products.push(product);
res.status(200).send(DB);
} else {
res.status(200).send(`Product already added to category.`);
};
} else {
res.status(200).send(`Category not found.`);
} });
Thanks in advance! Also sorry for the format.
There are a lot of repetitions, destructured variables that are not needed and syntactic errors in your code. Start by breaking down your problem into smaller chunks that are more manageable and then, test your endpoint with postman.
Let's start with your data and how it is structured:
var DB = [{
category: 'Pets',
products: [{
name: 'banjo',
color: 'grey',
mean: true,
description: "meows"
},
{
name: 'rigby',
color: 'black and white',
mean: false,
description: 'barks'
}
]
}]
// Lets say you want to add another pet into the pets category so in products
const obj = {
name: "frank",
color: "orange",
mean: false,
description: "glubs"
}
This is one of the things you could do to check if the object is not found and if not, you add it to your db.
DB.forEach((val) => {
if (val.products.name !== obj.name) {
DB[0].products.push(obj);
}
})
console.log(DB[0].products)
/**
*[{
name: 'banjo',
color: 'grey',
mean: true,
description: 'meows'
},
{
name: 'rigby',
color: 'black and white',
mean: false,
description: 'barks'
},
{
name: 'frank',
color: 'orange',
mean: false,
description: 'glubs'
}
]
*/
You postman request could look like this:
app.post("/product/add", (req, res) => {
// extract just what you need e.g name or ID if you have one...
const { name } = req.body;
// NOT efficient, best if you use a unique ID to look up
DB.forEach(value => {
if (value.products.name !== name) {
DB[0].products.push(req.body);
return res.status(200).send(JSON.stringify(DB));
}
else
return res.status(404).send("Not found or whatever");
});
});
OR using Array.prototype.findIndex you could do:
app.post("/product/add", (req, res) => {
const { name } = req.body;
const index = DB.forIndex(value => value === name);
if (index === -1) {
DB[0].products.push(req.body);
return res.status(200).send(JSON.stringify(DB));
}
else
return res.status(404).send("Not found or whatever");
}
});
Note: if your object name is the same than another one, your new object won't be pushed to the DB array. This would suggest you need to index your object with a unique identifier.
I have the following simple shema:
var userSchema = new Schema({
name : String,
age: Number,
_creator: Schema.ObjectId
});
var User = mongoose.model('User',userSchema);
What I want to do is create the new document and return to client, but I want to exclude the 'creator' field from one:
app.post('/example.json', function (req, res) {
var user = new User({name: 'John', age: 45, _creator: 'some ObjectId'});
user.save(function (err) {
if (err) throw err;
res.json(200, {user: user}); // how to exclude the _creator field?
});
});
At the end I want to send the new created user without _creator field:
{
name: 'John',
age: 45
}
Is it possible to make without extra find request to mongoose?
P.S:It's preferable to make it by
Another way to handle this on the schema level is to override toJSON for the model.
UserSchema.methods.toJSON = function() {
var obj = this.toObject()
delete obj.passwordHash
return obj
}
I came across this question looking for a way to exclude password hash from the json i served to the client, and select: false broke my verifyPassword function because it didn't retrieve the value from the database at all.
The documented way is
UserSchema.set('toJSON', {
transform: function(doc, ret, options) {
delete ret.password;
return ret;
}
});
UPDATE - You might want to use a white list:
UserSchema.set('toJSON', {
transform: function(doc, ret, options) {
var retJson = {
email: ret.email,
registered: ret.registered,
modified: ret.modified
};
return retJson;
}
});
Come across your question when I was trying to find a similar answer with pymongo. It turns out that in mongo shell, with the find() function call, you can pass a second parameter which specifies how the result document looks like. When you pass a dictionary with attribute's value being 0, you are excluding this field in all the document that come out of this query.
In your case, for example, the query will be like:
db.user.find({an_attr: a_value}, {_creator: 0});
It will exclude _creator parameter for you.
In pymongo, the find() function is pretty much the same. Not sure how it translate to mongoose though. I think it's a better solution compare to manually delete the fields afterwards.
Hope it helps.
I would use the lodash utilities .pick() or .omit()
var _ = require('lodash');
app.post('/example.json', function (req, res) {
var user = new User({name: 'John', age: 45, _creator: 'some ObjectId'});
user.save(function (err) {
if (err) throw err;
// Only get name and age properties
var userFiltered = _.pick(user.toObject(), ['name', 'age']);
res.json(200, {user: user});
});
});
The other example would be:
var _ = require('lodash');
app.post('/example.json', function (req, res) {
var user = new User({name: 'John', age: 45, _creator: 'some ObjectId'});
user.save(function (err) {
if (err) throw err;
// Remove _creator property
var userFiltered = _.omit(user.toObject(), ['_creator']);
res.json(200, {user: user});
});
});
You can call toObject() on the document to convert it to a plain JS object that you can freely modify:
user = user.toObject();
delete user._creator;
res.json(200, {user: user});
By following the MongoDB documentation, you can exclude fields by passing a second parameter to your query like:
User.find({_id: req.user.id}, {password: 0})
.then(users => {
res.status(STATUS_OK).json(users);
})
.catch(error => res.status(STATUS_NOT_FOUND).json({error: error}));
In this case, password will be excluded from the query.
font: https://docs.mongodb.com/v2.8/tutorial/project-fields-from-query-results/#return-all-but-the-excluded-field
I am using Mongoosemask and am very happy with it.
It does support hiding and exposing properties with other names based on your need
https://github.com/mccormicka/mongoosemask
var maskedModel = mongomask.mask(model, ['name', 'age']); //And you are done.
You can do this on the schema file itself.
// user.js
var userSchema = new Schema({
name : String,
age: Number,
_creator: Schema.ObjectId
});
userSchema.statics.toClientObject = function (user) {
const userObject = user?.toObject();
// Include fields that you want to send
const clientObject = {
name: userObject.name,
age: userObject.age,
};
return clientObject;
};
var User = mongoose.model('User',userSchema);
Now, in the controller method where you are responding back to the client, do the following
return res.json({
user: User.toClientObject(YOUR_ENTIRE_USER_DOC),
});