I'm trying to work around the fact that Datocms doesn't support a where filter in their GraphQL schema. Since there isn't that much data, I figured I could query all of it, and do the find on my end, but ... I'm not succeeding, at least not using "modern" methods.
What I get back when I query all of the data looks like this:
"foo": {
"data": {
"allGiveawayLandingPages": [
{
"lpSection": [
{},
{},
{},
{},
{},
{},
{},
{
"id": "34525949",
"products": [
{
"__typename": "PurchaseCardRecord",
"discountAmount": 50,
"discountAmountPct": null,
"discountEndDate": "2022-11-01T23:00:00+00:00",
"id": "44144096"
},
{
"__typename": "PurchaseCardRecord",
"discountAmount": null,
"discountAmountPct": null,
"discountEndDate": null,
"id": "44144097"
}
]
}
]
}
]
}
}
I need to find the object down in the "products" array by "id". This general question has been asked and answered lots of times, but the only answer I can get to work is from way back in 2013, and it seems to me there aught to be a more modern way to do it.
I'm doing this inside of a try/catch block, which I mention because it seems to be making this hard to debug (I'll come back to this):
export default async function createPaymentIntentHandler(req, res) {
const body = JSON.parse(req.body);
const {
productId,
productType
} = body;
let data;
if ('POST' === req.method) {
try {
switch (productType) {
case 'SeminarRecord':
data = await request({ query: singleSeminarQuery(productId) });
productObjName = 'seminar';
break;
default:
data = await request({ query: singleProductQuery(productId) });
productObjName = 'product';
}
/**
* Here's where I want to do my query / filtering
*/
// ... do more stuff and create Stripe paymentIntent
res.status(200).send({clientSecret: paymentIntent.client_secret})
} catch (error) {
logger.error({error}, 'Create Payment Intent error');
return res.status(400).end(`Create Payment Intent error: ${error.message}`);
}
} else {
res.status(405).end('Method not allowed');
}
}
My first, naive attempt was
const foo = await request({ query: ALL_PURCHASE_CARDS_QUERY });
const card = foo.data.allGiveawayLandingPages.find((page) => {
return page.lpSection.find((section) => {
return section?.products.find((record) => record.id === parentId)
})
});
logger.debug({card}, 'Got card');
In the abstract, aside from the fact that the above is fairly brittle because it relies on the schema not changing, I'd expect some similar sort of ES6 construction to work. This particular one, however, throws, but not in a particularly useful way:
[08:09:18.690] ERROR: Create Payment Intent error
env: "development"
error: {}
That's what I meant by it being hard to debug — I don't know why the error object is empty. But, in any case, that's when I started searching StackOverflow. The first answer which looked promising was this one, which I implemented as
...
const {
productId,
productType,
parentId
} = body;
...
function findCard(parent, id) {
logger.debug({parent}, 'searching in parent')
for (const item of parent) {
if ('PurchaseCardRecord' === item.__typename && item.id === id) return item;
if (item.children?.length) {
const innerResult = findCard(item.children, id);
if (innerResult) return innerResult;
}
}
}
if ('POST' === req.method) {
try {
...
const foo = await request({ query: ALL_PURCHASE_CARDS_QUERY });
const card = findCard(foo, parentId);
logger.debug({card}, 'Got card');
This similarly throws unhelpfully, but my guess is it doesn't work because in the structure, not all children are iterables. Then I found this answer, which uses reduce instead of my original attempt at find, so I took a pass at it:
const card = foo.data.allGiveawayLandingPages.reduce((item) => {
item?.lpSection.reduce((section) => {
section?.products.reduce((record) => {
if ('PurchaseCardRecord' === record.__typename && record.id === parentId) return record;
})
})
})
This is actually the closest I've gotten using ES6 functionality. It doesn't throw an error; however, it's also not returning the matching child object, it's returning the first parent object that contains the match (i.e., it's returning the whole "lpSection" object). Also, it has the same brittleness problem of requiring knowledge of the schema. I'm relatively certain something like this is the right way to go, but I'm just not understanding his original construction:
arr.reduce((a, item) => {
if (a) return a;
if (item.id === id) return item;
I've tried to understand the MDN documentation for Array.reduce, but, I don't know, I must be undercaffeinated or something. The syntax is described as
reduce((previousValue, currentValue) => { /* … */ } )
and then several variations on the theme. I thought it would return all the way up the stack in my construction, but it doesn't. I also tried
const card = foo.data.allGiveawayLandingPages.reduce((accumulator, item) => {
return item?.lpSection.reduce((section) => {
return section?.products.reduce((record) => {
if ('PurchaseCardRecord' === record.__typename && record.id === parentId) return record;
})
})
})
but the result was the same. Finally, not understanding what I'm doing, I went back to an older answer that doesn't use the ES6 methods but relies on recursing the object.
...
function filterCards(object) {
if (object.hasOwnProperty('__typename') && object.hasOwnProperty('id') && ('PurchaseCardRecord' === object.__typename && parentId === object.id)) return object;
for (let i=0; i<Object.keys(object).length; i++) {
if (typeof object[Object.keys(object)[i]] == 'object') {
const o = filterCards(object[Object.keys(object)[i]]);
if (o != null) return o;
}
}
return null;
}
if ('POST' === req.method) {
try {
...
const foo = await request({ query: ALL_PURCHASE_CARDS_QUERY });
const card = filterCards(foo);
logger.debug({card}, 'Got card');
This actually works, but ISTM there should be a more elegant way to solve the problem with modern Javascript. I'm thinking it's some combination of .find, .some, and .reduce. Or maybe just for ... in.
I'll keep poking at this, but if anyone has an elegant/modern answer, I'd appreciate it!
Related
I have state set as follow
const [stories, setStories] = useState([]);
I fetch Data from API in array, and i map the array and set the using setStories as:
setStories(prevState => prevState.concat({user: {name: 'XYZ', profile: 'ABC', stories: [{id: 1, image: 'testing'}];
The above codes are working fine, but i am stuck, when i have to concat the latest story if the id did not matched with fetched data. I have tried below solution but it didnot help:
stories.map(story => {
if(story && story.hasOwnProperty(key)){
//where above key is the user key fetched from the another API, i.e., user key
story?.[key].stories.map(storedStory =>
id(storedStory.id !== fetchedStory.id){
story?.[key].stories.concat({story})}
but the above code did not work, as it only mutate the state and is avoiding re-rendering.
Looking for a clean and efficient method to overcome this. THanks
It's hard to tell what you're trying to accomplish without seeing a full example. But I think your main problem is that you're not using the returned value from map, and from the naming it looks like you're appending the wrong element.
It will help to simplify first.
const newState = stories.map(story => {
if (story?.hasOwnProperty(key)) {
const found = story[key].stories.find(s => s.id === fetchedStory.id);
if (found) {
return story;
} else {
// Let's make a new object with the fetchedStory
// appended into THIS user's stories
return {
...story,
[key]: {
...story[key],
stories: [
...story[key].stories,
// This is supposed to be fetchedStory
// not `story` right??
fetchedStory,
]
}
}
}
} else {
return story;
}
});
setStory(newState);
Edit: You're having a hard time expressing your business logic, and the complexity of the data structure is not helping. So keep simplifying, encapsulate the complex syntax into functions then express your business logic plainly. Ie,
const appendStory = (originalObject, userId, storyToAppend) => {
return {
...originalObject,
[userId]: {
...originalObject[userId],
stories: [
...originalObject[userId].stories,
storyToAppend,
]
}
}
};
const userExistsInList = (users, user) => {
return users?.hasOwnProperty(user);
}
const newStoryAlreadyInStories = (stories, newStory) => {
return stories.find(s => s.id === newStory.id);
}
const newState = stories.map(story => {
if (userExistsInList(story, key)) {
const found = newStoryAlreadyInStories(story[key].stories, fetchedStory);
if (found) {
// User is already in state and the new story is already in the list
// Add business logic here
} else {
// User is already in state and the new story
// is not in their list
// Add business logic here
}
} else {
// User is not in the list yet
// Add business logic here
}
});
I am beating my head against a wall. I have updated to Apollo 3, and cannot figure out how to migrate an updateQuery to a typePolicy. I am doing basic continuation based pagination, and this is how I used to merged the results of fetchMore:
await fetchMore({
query: MessagesByThreadIDQuery,
variables: {
threadId: threadId,
limit: Configuration.MessagePageSize,
continuation: token
},
updateQuery: (prev, curr) => {
// Extract our updated message page.
const last = prev.messagesByThreadId.messages ?? []
const next = curr.fetchMoreResult?.messagesByThreadId.messages ?? []
return {
messagesByThreadId: {
__typename: 'MessagesContinuation',
messages: [...last, ...next],
continuation: curr.fetchMoreResult?.messagesByThreadId.continuation
}
}
}
I have made an attempt to write the merge typePolicy myself, but it just continually loads and throws errors about duplicate identifiers in the Apollo cache. Here is what my typePolicy looks like for my query.
typePolicies: {
Query: {
fields: {
messagesByThreadId: {
keyArgs: false,
merge: (existing, incoming, args): IMessagesContinuation => {
const typedExisting: IMessagesContinuation | undefined = existing
const typedIncoming: IMessagesContinuation | undefined = incoming
const existingMessages = (typedExisting?.messages ?? [])
const incomingMessages = (typedIncoming?.messages ?? [])
const result = existing ? {
__typename: 'MessageContinuation',
messages: [...existingMessages, ...incomingMessages],
continuation: typedIncoming?.continuation
} : incoming
return result
}
}
}
}
}
So I was able to solve my use-case. It seems way harder than it really needs to be. I essentially have to attempt to locate existing items matching the incoming and overwrite them, as well as add any new items that don't yet exist in the cache.
I also have to only apply this logic if a continuation token was provided, because if it's null or undefined, I should just use the incoming value because that indicates that we are doing an initial load.
My document is shaped like this:
{
"items": [{ id: string, ...others }],
"continuation": "some_token_value"
}
I created a generic type policy that I can use for all my documents that have a similar shape. It allows me to specify the name of the items property, what the key args are that I want to cache on, and the name of the graphql type.
export function ContinuationPolicy(keyArgs: Array<string>, itemPropertyKey: string, typeName: string) {
return {
keyArgs,
merge(existing: any, incoming: any, args: any) {
if (!!existing && !!args.args?.continuation) {
const existingItems = (existing ? existing[itemPropertyKey] : [])
const incomingItems = (incoming ? incoming[itemPropertyKey] : [])
let items: Array<any> = [...existingItems]
for (let i = 0; i < incomingItems.length; i++) {
const current = incomingItems[i] as any
const found = items.findIndex(m => m.__ref === current.__ref)
if (found > -1) {
items[found] === current
} else {
items = [...items, current]
}
}
// This new data is a continuation of the last data.
return {
__typename: typeName,
[itemPropertyKey]: items,
continuation: incoming.continuation
}
} else {
// When we have no existing data in the cache, we'll just use the incoming data.
return incoming
}
}
}
}
I'm new to react and as well to the terms of functional, imperative, declarative. And I get to know that pure function is easy to test. I am self taught to program with Javascript. So far, it is working but my goal is to learn to write clean and maintainable code.
my question is the method addProductToSaleList below is bad and untestable because it is imperative? and how can I do it differently.
class SaleComponent extends React.Component {
addProductToSaleList = (values, dispatch, props) => {
//filter product from productList
const productFound = props.productList.filter(product => {
if (values.productCode === product.code.toString()) {
return product
}
return undefined
})[0]
if (productFound) {
// filter sale list to check if there is already product in the list.
const detailFound = props.saleItem.details.filter(detail => {
if (productFound.name === detail.product) {
return detail
}
return undefined
})[0]
// if it is exist just increment the qty
if (detailFound) {
const { sub_total, ...rest } = detailFound
props.dispatcher('UPDATE_SALEDETAIL_ASYNC', {
...rest,
qty: parseInt(detailFound.qty, 10) + 1
})
// if it is not exist add new one
} else {
props.dispatcher('ADD_SALEDETAIL_ASYNC', {
product: productFound.id,
price: productFound.price,
qty: 1
})
}
} else {
alert('The product code you add is not exist in product list');
}
}
render() {
// Render saleList
}
}
I belive this question should go to Code Review, but I will give it a shot. Part of the code can be improved
const productFound = props.productList.filter(product => {
if (values.productCode === product.code.toString()) {
return product
}
return undefined
})[0]
First, filter function receives a callback and for each item that callback will be executed. If the callback returns a value interpreted as true, it will return the item in the new array the function will build. Otherwise, it will skip that item. Assuming you're trying to find one item in the code, you could use the function find which will return you that element directly (no need for [0]), or undefined if that item is not found. So your code could be rewrite to
const productFound = props.productList.find(product => values.productCode === product.code.toString());
Note: No IE support.
Then, if the value was not found, you could just alert and do an early return. (You might also want to handle errors differently, with a better format than plain alert).
The code would look like
if (!productFound) {
alert('The product code you add is not exist in product list');
return;
}
// rest of the function
in order to find details, you can use find method as well
const detailFound = props.saleItem.details.find(detail => productFound.name === detail.product);
and then just call the rest of the code
// if it is exist just increment the qty
if (detailFound) {
const { sub_total, ...rest } = detailFound
props.dispatcher('UPDATE_SALEDETAIL_ASYNC', {
...rest,
qty: parseInt(detailFound.qty, 10) + 1
})
// if it is not exist add new one
} else {
props.dispatcher('ADD_SALEDETAIL_ASYNC', {
product: productFound.id,
price: productFound.price,
qty: 1
})
}
Another improvement:
You're receiving a dispatch function as a parameter, but you're not using it. So you could remove it from function's declaration
(values, props) => { ... }
And you could split the last part into two different functions, something like
const getAction = details => `${detailFound ? 'UPDATE' : 'ADD'}_SALEDETAIL_ASYNC`;
const getObject = (details, productFound) => {
if (!details) {
return {
product: productFound.id,
price: productFound.price,
qty: 1
};
}
const { sub_total, ...rest } = detailFound;
return {
...rest,
qty: parseInt(detailFound.qty, 10) + 1
};
}
and then just call
props.dispatcher(getAction(details), getObject(details, productFound));
The end result would look like
addProductToSaleList = (values, props) => {
//filter product from productList
const productFound = props.productList.find(product => values.productCode === product.code.toString());
if (!productFound) {
alert('The product code you add is not exist in product list');
return;
}
// filter sale list to check if there is already product in the list.
const detailFound = props.saleItem.details.find(detail => productFound.name === detail.product);
const getAction = details => `${details ? 'UPDATE' : 'ADD'}_SALEDETAIL_ASYNC`;
const getObject = (details, productFound) => {
if (!details) {
return {
product: productFound.id,
price: productFound.price,
qty: 1
};
}
const { sub_total, ...rest } = details;
return {
...rest,
qty: parseInt(details.qty, 10) + 1
};
}
props.dispatcher(getAction(details), getObject(details, productFound));
}
my question is the method addProductToSaleList below is bad and
untestable because it is imperative
Well your code is testable, there are no external dependencies. So you could pass mocked values and props and add unit tests to that. That means, passing a fake values and props (they are just plain js object) and make assertions over that.
For instance:
You could mock dispatcher function and given the fake values in productList and saleItem.details you could see if dispatcher is called with the proper values. You should test different combinations of that
Mock alert function (Again, I would use another UI approach) and verify it is called, and that no other code is called (asserting that your fake dispatcher is not called). Something like this:
let actionToAssert;
let objectToAssert;
let values = { productCode: 'somecode' };
let props = {
productList: // your item listm with id and price, name, etc,
saleItem: {
details: // your details array here
}
dispatcher: (action, newObject) => {
actionToAssert = action;
objectToAssert = newObject;
}
}
addProductToSaleList(values, props); // make here assertions over actionToAssert and objectToAssert
Here is my code :
Search.prototype.makeQuery = function (data) {
let result = {};
if (data.orderId) {
result["order_id"] = data.orderId;
}
if (data.userMobileNumber) {
result["user.Number"] = {$regex : data.userMobileNumber}
}
if (data.userFullName) {
result["user.Name"] = {$regex: data.userFullName}
}
return result;
};
All I want is finding better way to optimize my code and reduce if condition in my code. Is there any suggestion ?
You can avoid the typing of if when you wrap it into a function and the typing of data with destructuring.
The advantage of wrapping the if in this case into a function is that you can simply test it, it is reusable and easy to read
Code
Search.prototype.makeQuery = function (data) {
let result = {}
let {orderId, userMobileNumber, userFullName} = data
setObjectValue(orderId, result, "order_id", orderId)
setObjectValue(userMobileNumber, result, "user.Number", {$regex : userMobileNumber})
setObjectValue(userFullName, result, "user.Name", {$regex: userFullName})
return result;
}
function setObjectValue(condition, object, key, value) {
if(condition) {
object[key] = value
}
}
Working Example
function makeQuery (data) {
let result = {}
let {orderId, userMobileNumber, userFullName} = data
setObjectValue(orderId, result, "order_id", orderId)
setObjectValue(userMobileNumber, result, "user.Number", {$regex : userMobileNumber})
setObjectValue(userFullName, result, "user.Name", {$regex: userFullName})
return result;
}
function setObjectValue(condition, object, key, value) {
if(condition) {
object[key] = value
}
}
let data = {
orderId: 1,
userMobileNumber: "016875447895",
userFullName: "John Doe"
}
let query = makeQuery(data)
console.log(query)
A simpler way:
Search.prototype.makeQuery = function (data) {
let result = {};
data.orderId && (result["order_id"] = data.orderId);
data.userMobileNumber && (result["user.Number"] = {$regex : data.userMobileNumber});
data.userFullName && (result["user.Name"] = {$regex: data.userFullName});
return result;
};
Let's imagine you have many fields or you want to modify them, you would create a map. Right now, your code works and my solution is overkill but it may be useful in the future:
const interestingData = new Map()
//I tried to imitate your values.
// Use whatever function you want here as a callback.
// value is the value you request, the callback must return the value you want to set.
interestingData.set("order_id", value => value)
interestingData.set("user.Number", value => ({ $regevalue: value }))
interestingData.set("user.Name", value => ({ $regevalue: value }))
//Tgis is a Factory in case you need several search.
const makeSearch = fields => data => {
let result = {}
fields.forEach((callBack, field) => {
if (data[field])
result[field] = callBack(data[field])
})
return result
}
//Creating a searching function
const myResearch = makeSearch(interestingData)
//Fake examples
const data1 = {
order_id: 'ertyui',
"user.Number": "ertyuio",
"user.Name": "ertyuio",
azerr: 123456
}
const data2 = {
order_id: 'ertyui',
"user.Number": "ertyuio",
}
console.log(myResearch(data1))
console.log(myResearch(data2))
It is not simpler but it is more extensible, and when you have many parameters, it is going to be much faster on a big scale. It is also reusable. Hope that helps!
Not sure if you'd consider this as code optimization, but you can get rid of the if statements using Object.assign:
Search.prototype.makeQuery = function (data) {
return Object.assign({},
data.orderId && { order_id: data.orderId },
data.userMobileNumber && {
'user.Number': { $regex : data.userMobileNumber },
},
data.userFullName && {
'user.Name': { $regex : data.userFullName },
},
)
};
If you can use newer JS features (with a transpiler or otherwise), you could use Object rest/spread for a slightly more concise syntax:
Search.prototype.makeQuery = (data) => ({
...data.orderId && { order_id: data.orderId },
...data.userMobileNumber && {
'user.Number': { $regex : data.userMobileNumber },
},
...data.userFullName && {
user.Name': { $regex : data.userFullName },
},
});
Edit 1: note that all these are pure functions, no mutations are taking place whatsoever
I'm using npm module traverse to filter data coming from mongodb / mongoose.
I might get this data:
[ { rating: 5,
title: { da: 'Web udvikling', en: 'Web Development' } },
{ rating: 5, title: { da: 'Node.js', en: 'Node.js' } } ]
'da' and 'en' indicates languages. I use traverse to filter mongoose data after current language like this:
var filtered = filterLanguage(lang, results);
// filter json obj by language
var filterLanguage = function(language, obj) {
return traverse(obj).map(function (item) {
if (this.key === language) {
this.parent.update(item);
}
});
};
I then show this in my template:
res.render('index', {
skills: filtered.skills
});
Finally I display it in the jade view:
ul.list-group
each skill, i in skills
if i < 5
li.list-group-item.sidebar-list-item= skill.title
Unfortunately it's displayed with quotes:
<ul>
<li>'Web Development'</li>
<li>'Node.js'</li>
</ul>
These quotes are not there in the unfiltered data (results.skill.title.da). So traverse is adding them. I used the module with 'plain' json and it's working perfectly.
The mongoose data seems plain and simple but of course there are a lot of properties on the prototype. Also traverse stalls if I don't omit '_id' (type bson/objectid) property from result set.
So traverse seems to have problems with mongoose data... Why is this? And how can I fix it?
-- EDIT --
I found a solution:
Before filtering I do this:
var json = JSON.parse(JSON.stringify(results));
var filtered = filterLanguage(lang, json);
This removes the quotes, but I'm not sure exactly what it does. Somehow converting the mongoose result to JSON? An explanation would be highly appreciated.
Fields in Mongoose documents are getters/setters, which seem to confuse either traverse or Jade/Pug.
The shortest method I found that seems to fix all of your issues is pretty ugly:
var filtered = filterLanguage(lang, results.map(r => JSON.parse(JSON.stringify(r))));
A more elaborate version:
var filtered = filterLanguage(lang, results.map(r => {
let j = r.toJSON()
j._id = j._id.toString()
return j;
}));
It would have been helpful to see what is the body of filterLanguage exactly or understand why it's called twice but as it stands, I don't think you need to use the traverse package at all.
A function such as below should do the trick and I even expanded it to work if the data is more tree-like and not as flat as represented in your example.
const reduceByLang = (data, lang) => {
// Look for a `lang` key in obj or
// if not found but still an object, recurse
const reduceByLangObj = (obj) => {
Object.keys(obj).forEach((key) => {
if (obj[key] === null) {
return;
}
if (obj[key][lang]) {
obj[key] = obj[key][lang]; // replace with desired lang
} else if (typeof obj[key] === 'object') {
reduceByLangObj(obj[key]); // recurse
}
});
return obj;
};
if (Array.isArray(data)) {
return data.map(reduceByLangObj);
} else {
return reduceByLangObj(data);
}
};
See example in JS Bin.
Also, if possible at all and if you do this type of selecting very often, I would look into saving the data in a different structure:
{ ratings: x, locales: { en: { title: 'Y' }, { da: { title: 'Z' } } } }
maybe, so that you can pick the selected language easily either in the query itself and/or in the controller.
EDIT: Checking for null.