…k with sidekiq
The problem is if the data has a ? in it, cassandra-cql will see that as a bind variable and throw an error. Have to pass it with the bind variables separate.
When creating a default_scope, we don't want to make a connection to Cassandra, because we may later fork, in which case the connection will be shared between processes, to bad effect. Instead, lazily initialize the connection the first time we need to make a call to Cassandra.
A row specification can take a DataSet as a value. In this case, fetch that data set, and then use it as inputs to the outer data set.
Borrowing the idea from Sequel, it makes sense to just consider a handle to a column group to be one case of a general scoped data set.