Please talk about performance implications over millions/ billions of rows.
Question asked by an attendee during Putting Complex Data Types to Work. @DeepakMajeti
Millions/billions of times the implications over a single row :-)
Kidding aside, it depends on what representation you choose. Flex tables, because the schema can change on each row and because the flex map is stored together, has a performance penalty over the columnar representation, until you decide to materialize frequently-used columns separately. Using the new complex types stuff, there isn't any overhead beyond the columnar representation, because it gets rewritten as columns, so it has the performance Vertica users have come to expect.
Can't find what you're looking for? Search the Vertica Documentation, Knowledge Base, or Blog for more information.