From the course: Advanced SQL for Data Science: Time Series

Unlock the full course today

Join today to access over 22,400 courses taught by industry experts or purchase this course individually.

Denormalizing time series data

Denormalizing time series data - SQL Tutorial

From the course: Advanced SQL for Data Science: Time Series

Start my 1-month free trial

Denormalizing time series data

- [Instructor] Sometimes we need to combine reference data with time series data. We can use denormalized tables for that. So the idea behind denormalizing is that we combine columns from multiple tables into a single table. This is called pre-joining, and some databases, like Postgres and Oracle and others, can support this kind of denormalizing using materialized views. With denormalized data, we do a join once, and this is useful when the data is queried multiple times. So rather than perform a join each time a query is executed, we can simply query the denormalized data table. So if query response time is more important than storage and preprocessing time, than denormalizing data is a good approach. So let's think about this in terms of the ingest pipeline. We have a data source, and it's producing time series data. So this could be, well, let's say CPU utilization and other server utilization metrics. We might want to add in related data, like the name of the department that's…

Contents