Now that we understand what time series data is and why we want to store it in a time series database, we’ve reached a new challenge. As with any application, we want to ensure our database queries are smart and performant, so let’s talk about how we can avoid some common pitfalls.
Indexing, the oft-recommended and rarely understood solution to all attempts at optimization, is applicable to most databases. Whether the time series database you’re using is built on Cassandra or MySQL or its own unique architecture, indexing affects your queries. Essentially, an index is a data structure that stores the values from a specific column, meaning that when we search by an indexed field, we have a handy shortcut to the values. When we search by unindexed fields, we have to discover the full path to the value, no shortcuts or magic tricks. Searching unindexed fields is like having to watch Frodo walk through Middle Earth unedited — it takes a long time.
Read more at The New Stack