PyMongo - Select object from an array having max value - mongodb

Question

A document in my DB looks like this :
{
"_id": ObjectId("5e92e63fad262707ff301d6c"),
"uknum": 30,
"area": "bath",
"ukelectors": 62355,
"ukresults": [
{
"party": "con",
"leader": "thatcher",
"ukvotes": 22544
},
{
"party": "lab",
"leader": "foot",
"ukvotes": 7259
},
{
"party": "sdp",
"leader": "jenkins",
"ukvotes": 17240
},
{
"party": "eco",
"leader": "whittaker",
"ukvotes": 441
}
]
}
Requirement :
I need to build a query in python to get the name of the party which won area: bath. Basically, check who got maximum votes and choose that party.
The idea was to use $max aggregation pipeline but it does not seem to work.

Answer 1

You can do that using either one of the aggregation-pipeline query :
Query 1 : Without use of $unwind, by using $reduce on array :
db.collection.aggregate([
{ $match: { area: "bath" } },
{
$addFields: {
ukresults: {
$let: {
vars: {
res: {
$reduce: {
input: "$ukresults",
initialValue: { votes: 0, party: {} },
in: {
votes: {
$cond: [
{ $gt: ["$$this.ukvotes", "$$value.votes"] },
"$$this.ukvotes",
"$$value.votes",
],
},
party: {
$cond: [
{ $gt: ["$$this.ukvotes", "$$value.votes"] },
"$$this",
"$$value.party",
],
},
},
},
},
},
in: "$$res.party",
},
},
},
},
]);
Test : MongoDB-Playground
Query 2 : With use of $unwind :
db.collection.aggregate([
{
$match: {
area: "bath"
}
},
{
$unwind: {
path: "$ukresults",
preserveNullAndEmptyArrays: true
}
},
{
$sort: {
"ukresults.ukvotes": -1
}
},
{
$limit: 1
}
])
Test : MongoDB-Playground
I would say these two might perform well as we're doing on mostly one document (Cause we've $match as first stage), maybe first query might take a wile to iterate if you've more elements in array, give it a try & choose one which helps most.
Ref : Check this pymongo documentation for aggregation examples : pymongo-aggregation.

Answer 2

The following aggregation query prints the ukresults sub-document with maximum votes.
The aggregation operator $max cannot be applied directly on the ukresults array, as the array has sub-documents rather than scalar values, like numbers. So, we use $reduce aggregation operator to extract the sub-document with maximum votes. Note that using the $reduce operator is a reduction operation on an array, so is using $max - the difference is the array element data type.
The PyMongo code:
import pymongo
import pprint
client = pymongo.MongoClient()
collection = client.test.testCollection
pipeline = [
{
"$match": { "area": "bath" }
},
{
"$addFields": {
"maxvotes": {
"$reduce": {
"input": "$ukresults",
"initialValue": { "ukvotes": 0 },
"in": {
"$cond": [
{ "$gt": [ "$$this.ukvotes", "$$value.ukvotes"] },
"$$this",
"$$value"
]
}
}
}
}
},
{
"$project": {
"_id": 0,
"area": 1,
"maxvotes": 1
}
}
]
pprint.pprint(list(collection.aggregate(pipeline)))
The output:
[{'area': 'bath',
'maxvotes': {'leader': 'thatcher', 'party': 'con', 'ukvotes': 22544.0}}]