Data Model

The Katmandoo database is built on a normalised relational data model that encapsulates the relationships between fundamental entities (for example, locations, years, operators, experimental units, treatments, etc) found in programs of planned research within the biosciences. Although initially created for use in plant improvement programs,  the core of the Katmandoo data model is based on generic principles and should prove adaptable in other settings.

A schematic of the storage of experiental results in Katmandoo is shown below. In this figure, parallelograms represent submodels within the data model, rectangles represent database tables and ellipses represent individual datum fields. Put simply,

• experimental units are identified geographically, temporally and contextually, and, in addition by relevent experiental design factors,
• treatments are considered a property of an experimental unit,
• a sampling (or observational) unit is some subdivision of the experimental unit and represents that experimental material on which measurements on characteristics (traits) of interest are recorded.
  

A schematic of the Version 2.1 data model showing the major submodels is given below. Each submodel consists of a set of logically related database tables, where a table is an abstraction of a physical or logical entity. For example, a geo-referenced site table relates to a physical place while a trial type table relates to a logical division of the experimantal program.

To maintain data integrity there is a strict sequence in which Katmandoo data structures must be populated. For quality control and data validation purposes, Katmandoo does not, in general, automatically add data to parent tables. The diagram provides some guidance in this regard but for specific cases see the Quick Start Guide.