MongoDB match Stage
The $match stage filters the documents entering the pipeline using the same query syntax as find(), and it is usually the most important stage for performance.
Filtering with $match
$match accepts a query document identical in syntax to what you would pass to find(): field-value equality checks, comparison operators like $gt and $lte, logical operators like $and and $or, and array or existence checks. Only documents that satisfy the query pass through to the next stage; everything else is discarded immediately.
A Simple Equality Match
db.orders.aggregate([
{ $match: { status: "completed" } }
])Using Query Operators Inside $match
Because $match reuses the standard query language, you can combine comparison and logical operators just as you would in a regular query, without needing any special aggregation syntax for this stage alone.
Combining Comparison and Logical Operators
db.orders.aggregate([
{ $match: {
status: { $in: ["shipped", "delivered"] },
total: { $gte: 100 },
createdAt: { $gte: ISODate("2026-01-01") }
} }
])This example keeps only orders that are shipped or delivered, worth at least 100, and created on or after the start of 2026. All three conditions must hold, since top-level fields in a query document are implicitly combined with AND.
$match Early vs. $match Late
A $match stage placed at the very start of a pipeline can use an existing index, exactly like a normal find() query would. A $match stage placed after a $group or $project, by contrast, filters on computed fields and cannot use an index because it is now operating on the pipeline's intermediate output rather than the raw collection.
Filtering Before and After Grouping
db.sales.aggregate([
{ $match: { region: "EMEA" } },
{ $group: {
_id: "$product",
totalRevenue: { $sum: "$amount" }
} },
{ $match: { totalRevenue: { $gt: 5000 } } }
])- Filters using the identical operators available to find(): $eq, $gt, $lt, $in, $exists, and more
- An early $match can use indexes, dramatically reducing work in later stages
- A later $match filters on computed values, useful for HAVING-style logic after $group
- Multiple $match stages are allowed and often clearer than one large compound condition
