A constraint on a global or central table controls what data can be inserted, updated or deleted from it. For example, a unique constraint on a set columns will prevent multiple rows from containing the same combination of values in those columns.
Primary key, unique key and foreign key constraints can be created on global and central tables.
A distribution constraint on a local table controls what join conditions must be applied to that table when it is joined with another local table at query time.
Note that a local table always contains an implicit primary key constraint on the message identity/state information, so that duplicate messages from the same source are discarded.