Query Latency
Query latency is the time between sending a query and receiving the results. In analytics it determines how responsive your dashboards feel and how quickly people can explore data and make decisions.
What drives query latency
Several things drive latency. How much data the query has to read is usually the biggest factor, which is why columnar storage and compression help so much: they cut the amount of data scanned. The efficiency of the query engine matters too, which is where vectorized execution earns its keep. And on large datasets, skipping irrelevant data through partition pruning and predicate pushdown is decisive.
High latency does not just annoy people. It changes behavior. When queries are slow, people stop exploring, dashboards get abandoned, and analysis that should take minutes takes a coffee break.
Low query latency over large data is the core promise of a good analytical database.
How Arc handles Query Latency
Arc keeps query latency low through columnar Parquet storage, ZSTD compression, a vectorized engine, and partition pruning, so large analytical scans return fast. Arrow IPC output reaches 9.2 million rows per second.
Related terms
Arc is a high-performance columnar database. Open Parquet on storage you own, single Go binary, production-ready in 30 seconds.