0
418
Viewsmatching on _id after $group stage aggregation
I have the following scenario in MongoDB:
Every record has its own _id and a parentId. If a parentId== "" then it's a true parent record. If the parentId has a value then that record is actually a child pointing to the parent record. The following shows one parent and its linked child.
{"_id": ObjectId('586c9d275d2f62e1634978db'), parentId="", count=1, <other fields>}
{"_id": ObjectId('586c9d275d2f62e163497811'), parentId=ObjectId('586c9d275d2f62e1634978db'), count=3, <other fields>}
I wanted a query where by I find all the parent records sorted by the count field where all parents and the child records are grouped together. For example, easiest explained via diagram:
ID6 has the highest count value which is associated to parent ID5. The next highest count is ID2 which is associated to parent ID1 finally ID4 is a parent and should also be returned so the results should be:
ID5, ID1, ID4
HoefMeistert helped me come up with the following query:
db.collection.aggregate(
[
{
$project: {
group_id : { $cond : { if: { $ne: [ "$parentId", "" ] }, then: "$parentId", else: "$_id" }},
count :1,
field1:1,
field2:1
}
},
{
$group: {
_id : "$group_id",
highest : { $max: "$count" }
},
"field1":{"$first":"$field1"},
"field2":{"$first":"$field2"},
},
{
$sort: {
highest : -1
}
}
]
);
The problem with this query is that it doesn't return the field1 and field2 associated to the parents I.e. ID1 and ID5 in the diagram. Is there a way to project the correct fields associated to the parents at the group stage? Otherwise if the group stage returns something like:
{'_id': ObjectId('586c9d275d2f62e1634978db'), 'highest': 2}
{'_id': ObjectId('586c9d0d5d2f62e1634978d5'), 'highest': 1}
{'_id': ObjectId('586c9d365d2f62e1634978e3'), 'highest': 0}
How can I rematch after the group to pull back the whole records for all the Ids above? I.e. 586c9d275d2f62e1634978db, 586c9d0d5d2f62e1634978d5, 586c9d365d2f62e1634978e3 ??
1 answers
Answer question0
Your query has an error, field1 and field2 need to be inside the $group dict:
db.collection.aggregate([
{
$project: {
group_id: { $cond: { if: { $ne: [ "$parentId", "" ] }, then: "$parentId", else: "$_id" }},
count: 1,
field1: 1,
field2: 1
}
},
{
$group: {
_id: "$group_id",
highest: { $max: "$count"},
field1: { "$first": "$field1"},
field2: { "$first":" $field2"},
},
},
{
$sort: {
highest : -1
}
}
]);
Result based on your diagram:
{ "_id" : "5", "highest" : 5, "field1" : ..., "field2" : ... }
{ "_id" : "1", "highest" : 3, "field1" : ..., "field2" : ... }
{ "_id" : "4", "highest" : 1, "field1" : ..., "field2" : ... }
Edit:
db.collection.aggregate([
{
$project: {
group_id: { $cond: { if: { $ne: [ "$parentId", "" ] }, then: "$parentId", else: "$_id" }},
count: 1,
field1: { $cond: { if: { $ne: [ "$parentId", "" ] }, then: null, else: "$field1" }},
field2: { $cond: { if: { $ne: [ "$parentId", "" ] }, then: null, else: "$field2" }},
}
},
{
$group: {
_id: "$group_id",
highest: { $max: "$count"},
field1: { "$max": "$field1"},
field2: { "$max":"$field2"},
},
},
{
$sort: {
highest : -1
}
}
]);
With this edit, in the group stage only the parents will have values for field1, field2, other documents will have null values. Than we can do a $max of them, to get the only value, the parent value.
Result will be the same like above, and field1, field2 will have values from parent documents