$group: การรวมข้อมูลและการคำนวณยอดรวม

The $group stage collapses multiple input documents into fewer output documents based on a grouping key. Documents that share the same key value are merged into a single output document, and accumulator operators compute aggregate values for each group. Think of it as MongoDB's equivalent of SQL's GROUP BY clause combined with aggregate functions like SUM and COUNT.

Every $group stage requires an _id field that defines the grouping key. The _id can be a field reference ('$field'), a computed expression, an object with multiple fields (for compound grouping), or null (to aggregate the entire collection into one document). The _id value in the output is the group key.

$sum is the most used accumulator. Pass 1 to count documents in each group, or a field reference to sum numeric field values. You can also pass an expression that evaluates to a number. $sum ignores non-numeric values and missing fields (they count as zero), making it safe to use on optional numeric fields.

$avg computes the arithmetic mean, $min finds the smallest value, and $max finds the largest value across all documents in a group. All three work on numeric values and also on strings and dates (for min/max). They ignore null and missing field values.

While you typically count by using { $sum: 1 } in a $group stage, MongoDB also provides a standalone $count stage to count the total number of documents in the stream. The $count stage emits a single document with the count under the field name you specify. It's equivalent to a $group with _id: null and a $sum: 1.

To group by multiple fields simultaneously, pass an object as the _id value. Each key in the object becomes part of the compound group key. The output documents have a nested _id object. This is the standard way to produce multi-dimensional aggregations like 'revenue by region and month'.

The $push accumulator collects all values from a field across the grouped documents into an array. This is useful for gathering all order IDs under a user, all tags under a category, or all user IDs who purchased a product. The result array can contain duplicates; use $addToSet instead to collect unique values.

$first and $last return the first and last field value encountered per group, in the order documents are processed. Because MongoDB does not guarantee document order unless you sort first, you should add a $sort stage before the $group stage to make $first/$last meaningful—for example, 'first order date per customer'.

The $group _id can be any expression, not just a field reference. You can group by a truncated date (year-week), a computed category, a mathematical bucket, or a substring of a field. This lets you create flexible analytical groupings without requiring pre-computed bucket fields in your stored documents.

Complex analytics often require two $group stages in sequence: the first groups at a detailed level (e.g., by user), and the second groups the first stage's results at a higher level (e.g., by region). This two-stage pattern avoids nested $group expressions and produces cleaner, more understandable pipelines.

Unlike $match, the $group stage cannot use an index—it must process all documents passed to it. This is why reducing the input with an early $match is critical. For very large datasets where $group uses too much memory, set allowDiskUse: true to spill to disk. Alternatively, pre-computing and storing aggregated values (the Computed Pattern) avoids runtime grouping on hot paths.

In this lesson you learned: $group collapses documents by a key defined in the _id field, accumulators like $sum, $avg, $min, $max, $push compute aggregate values per group, and two-stage grouping enables hierarchical analytics. Next up we explore $sort, $limit, and $skip to order and paginate aggregation results.