What should be done?

In $group aggregation Scalars dataset is skipped because aggregation grouped by value ({"_id", "$v"}) groups numbers of different types, and this causes output _id to be different number type between compat and target.

bson.D{{"$group", bson.D{
					{"_id", "$v"},
				}}},

It could be due to compat and target uses different order of insertion of documents.

Where?

See the issue number in the code

Definition of Done

• integration/compatibility tests added/updated;
• spot refactorings done;