How to use $set and dot notation to update embedded array elements using corresponding old element? - mongodb

Question

I have following documents in a MongoDb:
from pymongo import MongoClient
client = MongoClient(host='my_host', port=27017)
database = client.forecast
collection = database.regions
collection.delete_many({})
regions = [
{
'id': 'DE',
'sites': [
{
'name': 'paper_factory',
'energy_consumption': 1000
},
{
'name': 'chair_factory',
'energy_consumption': 2000
},
]
},
{
'id': 'FR',
'sites': [
{
'name': 'pizza_factory',
'energy_consumption': 3000
},
{
'name': 'foo_factory',
'energy_consumption': 4000
},
]
}
]
collection.insert_many(regions)
Now I would like to copy the property sites.energy_consumption to a new field sites.new_field for each site:
set_stage = {
"$set": {
"sites.new_field": "$sites.energy_consumption"
}
}
pipeline = [set_stage]
collection.aggregate(pipeline)
However, instead of copying the individual value per site, all site values are collected and added as an array. Intead of 'new_field': [1000, 2000] I would like to get 'new_field': 1000 for the first site:
{
"_id": ObjectId("61600c11732a5d6b103ba6be"),
"id": "DE",
"sites": [
{
"name": "paper_factory",
"energy_consumption": 1000,
"new_field": [
1000,
2000
]
},
{
"name": "chair_factory",
"energy_consumption": 2000,
"new_field": [
1000,
2000
]
}
]
},
{
"_id": ObjectId("61600c11732a5d6b103ba6bf"),
"id": "FR",
"sites": [
{
"name": "pizza_factory",
"energy_consumption": 3000,
"new_field": [
3000,
4000
]
},
{
"name": "foo_factory",
"energy_consumption": 4000,
"new_field": [
3000,
4000
]
}
]
}
=> What expression can I use to only use the corresponding entry of the array?
Is there some sort of current-index operator:
$sites[<current_index>].energy_consumption
or an alternative dot operator (would remind me on difference between * multiplication and .* element wise matrix multiplication)?
$sites:energy_consumption
Or is this a bug?
Edit
I also tried to use the "$" positional operator, e.g. with
sites.$.new_field
or
$sites.$.energy_consumption
but then I get the error
FieldPath field names may not start with '$'
Related:
https://docs.mongodb.com/manual/reference/operator/aggregation/set/#std-label-set-add-field-to-embedded
In MongoDB how do you use $set to update a nested value/embedded document?

Answer 1

If the field is member of an array by selecting it you are selecting all of them.
{ar :[{"a" : 1}, {"a" : 2}]}
"$ar.a" = [1 ,2]
Also you cant mix update operators with aggregation, you cant use things like
$sites.$.energy_consumption, if you are doing aggregation you have to use aggregate operators, with only exception the $match stage where you can use query operators.
Query
alternative slightly different solution from yours using $setField
i guess it will be faster, but probably little difference
no need to use javascript it will be slower
this is >= MongoDB 5 solution, $setField is new operator
Test code here
aggregate(
[{"$set":
{"sites":
{"$map":
{"input":"$sites",
"in":
{"$setField":
{"field":"new_field",
"input":"$$this",
"value":"$$this.energy_consumption"}}}}}}]
)

Answer 2

use $addFields
db.collection.update({},
[
{
"$addFields": {
"sites": {
$map: {
input: "$sites",
as: "s",
in: {
name: "$$s.name",
energy_consumption: "$$s.energy_consumption",
new_field: {
$map: {
input: "$sites",
as: "value",
in: "$$value.energy_consumption"
}
}
}
}
}
}
}
])
mongoplayground

Answer 3

I found following ugly workarounds that set the complete sites instead of only specifying a new field with dot notation:
a) based on javascript function
set_stage = {
"$set": {
"sites": {
"$function": {
"body": "function(sites) {return sites.map(site => {site.new_field = site.energy_consumption_in_mwh; return site})}",
"args": ["$sites"],
"lang": "js"
}
}
}
}
b) based on map and mergeObjects
set_stage = {
"$set": {
"sites": {
"$map": {
"input": "$sites",
"in": {
"$mergeObjects": ["$$this", {
"new_field": "$$this.energy_consumption_in_mwh"
}]
}
}
}
}
}
If there is some kind of $$this context for the dot operator expression, allowing a more elegant solution, please let me know.