Aggregate $match _id $eq $toObjectId not working

[David_Camps]

Using Compass 1.30.1, I was testing an aggregation and getting unexpected results. A $match was not working as expected. The simplified aggregation is:

db.getCollection('dlsComponents').aggregate([
  { $match: { library: 'Library1', collection: 'Collection1', media: 'Images', object: 'Image3' } }
 ])

And this gives the expected result by finding a document:

{ _id: ObjectId("61fc458b46d7874a3a97ef79"),
  library: 'Library1',
  collection: 'Collection1',
  media: 'Images',
  object: 'Image3',
  info: 'Image: 1/1/Images/Image3 Info', …

try to get the same document by _id:

db.getCollection('dlsComponents').aggregate([
  { $match: { _id: { $eq: { $toObjectId: "61fc458b46d7874a3a97ef79" } } } }
 ])

does not find a document. Why does the second $match not find a document?

[David_Camps]

I found that:

{ $match: { $expr: { $eq: [ ‘$_id’, ‘$$imageId’ ] } } }

does work ($$imageId is an ObjectId used in the non-simplified aggregate). Maybe the { $eq: ‘$value’ } format does not work in pipelines.

[jbalint (Jess Balint)]

Hi David,

The $eq used in find()/$match (without $expr) must specify an exact value: https://docs.mongodb.com/manual/reference/operator/query/eq/

You can use $eq: ObjectId("...")

Jess

[steevej (Steeve Juneau)]

Your last post made me think that may be $toObjectId works only inside $expr. I tried

{ $match: { $expr : { $eq: [ '$_id' , { $toObjectId: "61fc458b46d7874a3a97ef79" } ] } } }

and it works.

[steevej (Steeve Juneau)]

See on a related topic

$lookup foreignField is ObjectId

That is why it is better to keep your ObjectId as ObjectId rather than strings. In addition an ObjectId probably takes less space than it’s string representation. But your you’re in luck, see the following: The drawbacks is that each and every string must be converted. Most likely no index can be used because of the conversion. You might need an alternate form of $lookup because you might need to use let and may be enclosed the conversion into an $expr.

and

Aggregation between 2 tables with join using field that is array of strings

Your problem is that you keep your document ids as string in your batches collection. This means that every time you lookup you must convert your string ids into ObjectId or the other way around. This means you will not be able to use the index on the converted values. Since it is best to keep ObjectId as ObjectId, the first step is to make sure DocumentIDs holds the ID as ObjectId. Recent $lookup is more powerfull than the one you shown. See https://docs.mongodb.com/manual/reference/operat…

[Crown_International_Technology_Pvt_Ltd_CIT]

even gpt4 doesn’t solve my problem, thank

[steevej (Steeve Juneau)]

Did you told GPT4 what was your problem? Or did you do like you did here., just saying that you have a problem.