Colocated column
Colocated columns are an alternative to creating new dataset copies when adding shard keys. When you want to establish a new shard key, you usually need to create a new dataset copy. Because each new dataset copy creates a complete copy of the underlying dataset, this linearly increases the storage required in your cluster. Colocated columns allow you to avoid creating new dataset copies when adding shard keys, if the shard key fulfills certain criteria.
If the shard key you want to add has a one-to-one or many-to-one relationship with an existing shard key, you can create a colocated column instead of creating a new copy of your dataset. The important point is that each instance of the new shard key must map to exactly one instance of an existing shard key.