The following assumption is made:

• The datasets will always be able to fit in memory

Run the following to start the old database in a docker container:

make db/old

Run the following to start the new database in a docker container:

make db/new

Run the following to diff the two databases. The output will be written to diff.edn. Output will be a sequence of maps with the account id and the associated status :missing, :new, or :corrupt

make db/diff

To run the tests:

make test