...

SWS Documentation on Compilation of aggregates: https://sws.fao.org/doc/docs/statisticalprocesses/cross-cutting/aggregation/

Entities

Domains. Thematic (and institutional) datasets. Domains are subject areas such as "TCL. Crops and livestock products".

Elements. Phenomena observed for a given domain. E.g., Yield, Stocks for TCL. 

Items. These constitute the specific entities used to observe an element. They correspond to the individual measurements in a dataset. For example, relevant items for the element TCL are, among others, "Wheat" and "Rice". Items are usually defined in statistical classifications (e.g., the Central Product Classification, CPC, maintained by UNSD), but may also be defined ad-hoc for the purpose of data collection or processing. Itemsare defined, at least, by a label and a code. Usually, items are hierarchically organized, but they may also be defined in lists or in faceted structures.

Aggregates are set of items. The hierarchical structure provided in an official statistical classification may be used for data aggregation and other, ad-hoc aggregates may be defined for data processing and/or data dissemination purposes. When new aggregates are defined, they are given “new” codes, not present in any official or public statistical classification. E.g., For TCL, "Cereals, primary"