Joining large database tables

A computer product comprising a computer readable medium storing a plurality of instructions for controlling a plurality of computer nodes to perform a join of a plurality of tables distributed across a plurality of computer nodes of a network, the instructions comprising:

• receiving, at each of the plurality of computer nodes, a query identifying the plurality of tables to be joined;
• for a first table of the plurality of tables:
• creating, by each of the plurality of computer nodes, a first active flag structure that identifies any table entries satisfying one or more criteria and that are stored on the computer node;
• identifying, by a first of the plurality of computer nodes, a first set of first entries in a first table that are identified as satisfying the one or more criteria by the first active flag structure, the first active flag structure including first flags operable to indicate which first entries are to be used in an output table;
• iterating, by the first computer node, over the first set of first entries in the first table to identify keys corresponding to a second set of second entries in a second table, the second table being one of the plurality of tables;
• using the identified keys to generate a request corresponding to the second set of entries in the second table;
• updating, by the first computer node, at least one of the first flags of the first active flag structure to be false based on a response to the request;
• subsequent to the updating, identifying a first group of first flags of the first active flag structure that are true;
• and
• retrieving data corresponding to the first group of first flags for creating the output table.