Right now a query like the following fails to parse:

T1 = load("https://uwdb.s3.amazonaws.com/sampleData/TwitterK.csv",
csv(schema(a:int,b:int),skip=0));
store(T1, TwitterK, [a, b]);
T2 = scan(TwitterK);
sink(T2);

MyriaError: Error 404 (Not Found): That dataset was not found

RACO is trying to look up the relation TwitterK in the catalog even though it doesn't yet exist. Perhaps we need to introduce some sort of boundary in the Sequence logical operator to denote when a relation name can be considered eligible for validation.

This is a problem for any queries that need to materialize intermediate results (e.g., to push operations into Postgres).