Select Language:
If you’re thinking about using AWS ElastiCache with Valkey 8.2, it’s important to know that it doesn’t only support vector search. While vector search is a new feature in Valkey 8.2, it actually builds on the capabilities of the previous version, Valkey 8.1. This means ElastiCache v8.2 works smoothly with Valkey v8.1, and you can still make use of other search options.
In fact, ElastiCache Valkey 8.2 allows for different types of searches, not just vector. When you create an index, it supports three kinds of data fields:
– Vector fields — used for similarity searches based on vectors
– Number fields — used for range searches
– Tag fields — used for filtering based on tags
Number fields are particularly useful for range searches and can be used alone or alongside vector searches. However, there’s an important rule to keep in mind: when setting up an index, you must include at least one vector field. You can’t create an index that only contains metadata, like number or tag fields, without any vector components.
Once you have an index with at least one vector field, you’ll be able to perform searches that involve range queries and filtering on number and tag fields. The syntax for vector search, which uses the “=>” operator with KNN (K-Nearest Neighbors), is specific to finding similar vectors. But behind the scenes, your search can still filter results based on other field types.
This requirement to include at least one vector field in the index is part of how Valkey 8.2’s vector search is designed. It’s a constraint that exists even if your main goal is to perform metadata searches without heavy reliance on vector similarity.
For more details, you can check out the official documentation on AWS’s website, which explains the features and limits of vector search in ElastiCache, how engine versions work, and an overview of vector search capabilities.
ChatGPT
Perplexity
Gemini AI
Grok AI
