#Materialize in a sentence update#
In contrast to the traditional approach, Materialize let teams continually update the answers to their queries as new data comes in. It’s also slow and doesn’t allow teams to nimbly make decisions. This once-a-day update cycle works, but it’s inopportune to continually recompute an answer that you once knew. Instead, you want to compute the answer once and then cache it until you derive a new answer. These kind of computations are therefore relatively expensive, so it’s not feasible to execute them on the fly. This approach makes intuitive sense: the working set is often larger than any machine’s memory and you need to process all of it. Traditionally, to get answers, some batch processing system performs a query over all of the data every night.
You might have a ton of data that comes in all day, and each team in your org wants to answer some questions using that data. It’s useful to think of these questions as the kind of queries you might use to power business intelligence dashboards. Materialize lets teams very quickly get answers to questions they routinely ask. It takes to refresh answers to common queries, Materialize can make that happen.įor a sense of scale, it can take queries that most teams would run once-per-dayĪnd instead provide sub-second or single-digit second answers. If you perform any OLAP queries over relational data and want to reduce the time With complex aggregations), and can do incremental updates in the presence ofĪrbitrary inserts, updates, and deletes in the input streams. Views than is common in traditional databases (e.g. Importantly, Materialize supports incrementally updating a much broader set of It’s asked, Materialize continually updates the answer and gives you the So, rather than recalculating the answer each time Materialize instead keeps the results of the queries and incrementally updates Question before, which means it can take a long time to come up with anĪnswer, each and every time you pose the query. Why not just use your database’s built-in functionality to perform these sameĬomputations? Because your database often acts as if it’s never been asked that Materialize lets you ask questions about your data, and then get low-latency,Ĭorrect answers, even as the underlying data changes. If that jargon-filled sentence doesn’t answer all of your questions, we’ll cover some important conceptual topics related to Materialize:
Materialize is a streaming SQL materialized view engine.