The reason for this is the generator in Catmandu::Store::DBI::Iterator,
that fetches the records in batches: for every batch a query is executed,
using "order by id", "offset" and "limit". This means that the underlying
dbi system has to jump further every time it is accessed.
The reason for this is the generator in Catmandu::Store::DBI::Iterator, that fetches the records in batches: for every batch a query is executed, using "order by id", "offset" and "limit". This means that the underlying dbi system has to jump further every time it is accessed.
Better to cache the cursor