Run query based on output from previous query.

Rework the executor and chaining model so we can run multiple queries in the same line. The second query should be able to take arguments from the produced result set, for instance:

SELECT ... | ... | exec INSERT INTO foo SET x=$1, y=$2

So each column would be mapped into a numbered or named parameter ($1, $2 etc) and the second query would be executed once for every row in the previous step. This process should be repeatable.