Mongodb refers in query - mongodb

Question

I'm newbie in mongo, so:
I have a collection 'my_collection' like this:
{
"_id" : ObjectId("5546329a470000850084a621"),
"company_name" : "Microsoft",
"type" : 'company',
}
{
"_id" : ObjectId("5546329a470000850084a622"),
"company_name" : "Google",
"type" : 'company',
}
{
"_id" : ObjectId("5546329a470000850084a623"),
"company_name" : "Apple",
"type" : 'company',
}
{
"_id" : ObjectId("5546329a470000850084a624"),
"name" : "John",
"surname" : "Smith",
"type" : 'person',
}
{
"_id" : ObjectId("5546329a470000850084a625"),
"name" : "Eugene",
"surname" : "Harper",
"type" : 'person',
"company_id" : '5546329a470000850084a622',
}
{
"_id" : ObjectId("5546329a470000850084a626"),
"name" : "Philipp",
"surname" : "Hoffman",
"type" : 'person',
"company_id" : '5546329a470000850084a622',
}
{
"_id" : ObjectId("5546329a470000850084a627"),
"name" : "Adam",
"surname" : "Jonson",
"type" : 'person',
}
{
"_id" : ObjectId("5546329a470000850084a628"),
"name" : "Bruse",
"surname" : "Willis",
"type" : 'person',
"company_id" : '5546329a470000850084a623',
}
I need to get a list of companies that have employees. (ie if an employee refers to a company, it must be listed)
I think in MySQL my query would be something like this:
SELECT * from my_collectionWHERE type = company AND (type = person AND company_id = _id)
So, in the result I expect two companies:
1. Google
2. Apple
What query it needed in mongo to get this result?
{
"_id" : ObjectId("5546329a470000850084a621"),
"company_name" : "Microsoft",
}
{
"_id" : ObjectId("5546329a470000850084a622"),
"company_name" : "Google",
}
It's wrong
$collection = $db -> my_collection->find(
array(
"type" => 'company',
"_id" => { $in: [
"type" => 'person',
"_id" => ???,
]
}
)
);

Answer 1

Your both the ids are different. So you are using $in is perfect. I don't know much about php but $in query takes an array of argument as parameters.
So pass ["5546329a470000850084a621", "5546329a470000850084a622"] value in $in query.
Your code will be something like this:
$collection = $db -> my_collection->find(
array(
"type" => 'company',
"_id" => { $in: ["5546329a470000850084a621", "5546329a470000850084a622"]
}
)
);

Answer 2

If it's not possible to convert the schema then you would need to run two queries; one to get the company ids for documents which have a person "type" with the result mapped to ObjectIds and the final to query the collection with the above ids.
Consider the following queries:
const company_ids = db.my_collection.distinct("company_id",
{ "type": "person", "name": { "$exists": true } }
).map(ObjectId);
db.my_collection.find({ "_id": { "$in": company_ids } }, { "company_name": 1 });
For an effective query which leverages the $lookup operator that can do a self-join on the same collection, you need to change your schema first i.e. cast the company_id string to ObjectId.
A couple of ways you can go about the schema conversion. For relatively large collections use the bulkWrite() method with batched updates as:
let bulkUpdateOps = [],
cursor = db.my_collection.find({ "company_id": { "$type": 2 } });
cursor.forEach(doc => {
const { _id, company_id } = doc;
let temp = new ObjectId(company_id);
bulkUpdateOps.push({
"updateOne": {
"filter": { _id },
"update": { "$set": { company_id: temp } },
"upsert": true
}
});
if (bulkUpdateOps.length === 1000) {
db.my_collection.bulkWrite(bulkUpdateOps);
bulkUpdateOps = [];
}
});
if (bulkUpdateOps.length > 0) {
db.my_collection.bulkWrite(bulkUpdateOps);
}
or for small collections as straightforward as
db.my_collection.find({ "company_id": { "$type": 2 } }).forEach(doc => {
doc.company_id = new ObjectId(doc.company_id);
db.test.save(doc);
});
After the schema conversion, you can then query using the aggregation framework which has a $lookup pipeline step that you can use to create a "self-join" of the collection, a final $match stage to query the companies with the employees and return the desired result.
Consider running the following aggregate operation:
db.my_collection.aggregate([
{
"$lookup": {
"from": "my_collection",
"localField": "_id",
"foreignField": "company_id",
"as": "employees"
}
},
{ "$match": { "employees.0": { "$exists": true } } }
])
Sample Output
/* 1 */
{
"_id" : ObjectId("5546329a470000850084a622"),
"company_name" : "Google",
"type" : "company",
"employees" : [
{
"_id" : ObjectId("5546329a470000850084a625"),
"name" : "Eugene",
"surname" : "Harper",
"type" : "person",
"company_id" : ObjectId("5546329a470000850084a622")
},
{
"_id" : ObjectId("5546329a470000850084a626"),
"name" : "Philipp",
"surname" : "Hoffman",
"type" : "person",
"company_id" : ObjectId("5546329a470000850084a622")
}
]
}
/* 2 */
{
"_id" : ObjectId("5546329a470000850084a623"),
"company_name" : "Apple",
"type" : "company",
"employees" : [
{
"_id" : ObjectId("5546329a470000850084a628"),
"name" : "Bruse",
"surname" : "Willis",
"type" : "person",
"company_id" : ObjectId("5546329a470000850084a623")
}
]
}