The Analytics development team has done something very clever when building an API-to-AMT implementation: make use of the link elements shown in the screenshot below. This is a problem for Project Meadowlark - Exploring Next Generation Technologies because the Meadowlark does not (as of ) respond with these links. Ideally, any "data out" scenario from an Ed-Fi API should be built so that it would work with anyone's implementation of an Ed-Fi API.
API to AMT Algorithm
...
Another option is for the AMT data lake solution to create a “lookup table” that stores both the link.href value and a calculated hash value. In SQL terms this would be like having the following tables and joins.
This has the advantage of being useful for Meadowlark. The “IdMappings” table is quick-and-dirty and can probably down to just “href” and “id”. In fact, this is similar to how Meadowlark handles referential integrity checks, through the help of a References lookup table. Side note: the lookup table in the middle of the diagram probably does not need source and destination, since the href to a student would be the same regardless of whether the source is studentEducationOrganizationAssociation or any other document that references a student. And the "destination" (students) is already embedded into the href value and thus easy for a human to understand.
...