Mongo Distinct Index for Dot Notated - mongodb

Question

I have a mongodb v4.2.2 hosted on Atlas. Have a collection
objects and two fields in it: metadata as array of objects, each
object has one field key AND second field named as model. Added compound index
model->metadata.key as regular one. While running db command
db.objects.explain('executionStats').distinct('metadata.key',
{model: ObjectId('5e18aff58a5aaffdc3d6f26d')}) from console it
doesn't use my model->metadata.key index and doesn't use DISTINCT_SCAN.
But if I move data to another collection lets say objectKeys that
has two fields like key, model and again add regular compound
index model->key and run command
db.objectKeys.explain('executionStats').distinct('key', {model:
ObjectId('5e18aff58a5aaffdc3d6f26d')}) it is successfully applying DISTINCT_SCAN
and use the index.
Question. How to force MongoDB to use indexes for dot notated
field during distinct operation?
UPDATE DETAILS:
1.
db.objects.getIndexes();
{
"v" : 2,
"key" : {
"model" : 1,
"metadata.key" : 1,
"metadata.value" : 1
},
"name" : "model_1_metadata.key_1_metadata.value_1",
"ns" : "my_db.objects",
"background" : true
}
db.objects.explain('executionStats').distinct('metadata.key', {model: ObjectId('5e18aff58a5aaffdc3d6f26d')})
"executionStages" : {
"stage" : "FETCH",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 1,
"advanced" : 0,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"docsExamined" : 0,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 1,
"advanced" : 0,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"keyPattern" : {
"model" : 1,
"metadata.key" : 1,
"metadata.value" : 1
},
"indexName" : "model_1_metadata.key_1_metadata.value_1",
"isMultiKey" : true,
"multiKeyPaths" : {
"model" : [ ],
"metadata.key" : [
"metadata"
],
"metadata.value" : [
"metadata"
]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"model" : [
"[ObjectId('5e18aff58a5aaffdc3d6f26d'), ObjectId('5e18aff58a5aaffdc3d6f26d')]"
],
"metadata.key" : [
"[MinKey, MaxKey]"
],
"metadata.value" : [
"[MinKey, MaxKey]"
]
},
"keysExamined" : 0,
"seeks" : 1,
"dupsTested" : 0,
"dupsDropped" : 0
}
}
2.
db.object_keys.getIndexes();
{
"v" : 2,
"key" : {
"models" : 1,
"key" : 1
},
"name" : "models_1_key_1",
"ns" : "my_db.object_keys",
"background" : true
},
db.object_keys.explain('executionStats').distinct('key', {models: ObjectId('5e18aff58a5aaffdc3d6f26d')})
"executionStages" : {
"stage" : "PROJECTION_COVERED",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 1,
"advanced" : 0,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"transformBy" : {
"_id" : 0,
"key" : 1
},
"inputStage" : {
"stage" : "DISTINCT_SCAN",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 1,
"advanced" : 0,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"keyPattern" : {
"models" : 1,
"key" : 1
},
"indexName" : "models_1_key_1",
"isMultiKey" : true,
"multiKeyPaths" : {
"models" : [
"models"
],
"key" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"models" : [
"[ObjectId('5e18aff58a5aaffdc3d6f26d'), ObjectId('5e18aff58a5aaffdc3d6f26d')]"
],
"key" : [
"[MinKey, MaxKey]"
]
},
"keysExamined" : 0
}
}
So as you can see in the first case it uses IXSCAN means index scan and in the second case it uses DISTINCT_SCAN that is considered more faster. The differences is that in first case the metadata field is array of objects and in the second case the field key is string