we create a set with an identifier, xx-NO or xx-code.
　There may be some "partition" codes in a set, and the set may be divided into some basic classes,
　　i.e. subsets.

　We have data of "business office" with the identifier of "BO-NO", "business office" is an entity, and it's describe as noted below:

　Business Office:｛BO-NO, BO-name,...｝[ R ]

　If and when there is, in the "business office" set, a "partition" code of "classification by location" with the value of "1" for "domestic" and "2" for "subsidiary abroad", then the entity must be divided into two subsets as noted below:

　Business Office:｛BO-NO, BO-name, by location,...｝[ R ]
　　｜
　　＝ by location
　　｜＿＿＿＿ domestic
　　｜
　　｜＿＿＿＿ subsidiary abroad

　It must be explicitly declared by which rule--a "partition" code"-- is applied to create subsets, like "by-location" as in example above.