Will aggregation pipeline use index for derived field? - mongodb

Question

In aggregation pipeline I have this:-
{
"$project" => {
account_name_i: { "$toLower" => "$account_name" },
}
}
{
"$sort" => {
account_name_i: 1
}
}
and I have index { account_name: 1 }
My Question is will $sort use index on account_name? If no, there is any other way we can achieve this in aggregation pipeline?

Answer 1

No. The aggregation pipeline can only use a standard index on a $match or $sort phase that's at the beginning of a pipeline. The rules for using indexes with aggregation pipelines are described in detail in the manual.
You aim to sort on the lower-case version of account_name_i, most likely to function as a case-insensitive sort. To achieve this with an index, store a lower-case-normalized version of account_name_i in each document
{ "account_name_i" : "TruMAn's HABerdaShEry", "account_name_i_lc" : "truman's haberdashery" }
and index the normalized field ({ "account_name_i_lc" : 1 }).

Answer 2

If you are using Aggregation Pipeline Stages like $group or $match on a collection then it will
produce a new set of documents from that collection.
as it is new set of documents which means no index is defined for it and hence we cant use the index in Aggregation Pipeline
Note: although if you use $sort as first stage of Aggregation Pipeline then it will utilize the index of
the collection as input for this stage is collection itself.

Answer 3

Kind of, though only with preprocessing. You can do an output after your project stage and query the output collection for sorting:
db.yourSource.aggregate([
{
"$project" => {
account_name_i: { "$toLower" => "$account_name" },
}
},
{
$out: "intermediate"
}
])
db.intermediate.ensureIndex({account_name_i:1})
db.intermediate.find({}).sort({account_name_i:1})