dot Stop testing, start deploying your AI apps. See how with MIT Technology Review’s latest research.

Download now

7.2 Sorted Indexes

back to home

7.2 Sorted Indexes

In the previous section, we talked primarily about searching, with the ability to sort results by referencing data stored in HASHes. This kind of sorting works well when we have a string or number that represents the actual sort order we’re interested in. But what if our sort order is a composite of a few different scores? In this section, we’ll talk about ways to combine multiple scores using SETs and ZSETs, which can offer greater flexibility than calling SORT.

Stepping back for a moment, when we used SORT and fetched data to sort by from HASHes, the HASHes behaved much like rows in a relational database. If we were to instead pull all of the updated times for our articles into a ZSET, we could similarly order our articles by updated times by intersecting our earlier result SET with our update time ZSET with ZINTERSTORE, using an aggregate of MAX. This works because SETs can participate as part of a ZSET intersection or union as though every element has a score of 1.