ContextQMD
Back to Redis

Redis 6.2

content/develop/whats-new/6-2.md

latest3.7 KB
Original Source

Redis version 6.2 introduces new capabilities designed to improve data indexing, querying, and analytics. This update brings multi-value indexing, expanded wildcard query support, and a new probabilistic data structure for quantile estimation. Additionally, significant enhancements to Redis streams and time series data processing offer greater flexibility for developers working with real-time and historical datasets. Over 25 new commands have been added to Redis that address key feature requests and further extending its capabilities. Below is a detailed breakdown of these improvements.

New features

Multi-value indexing and querying

Redis now supports indexing and querying multi-value attributes across all field types, including TEXT, TAG, NUMERIC, GEO, and VECTOR. Developers can define JSONPath expressions leading to arrays or multiple scalar values, overcoming the previous limitation of indexing only single scalar attributes.

Wildcard query support

Redis Search now enables suffix and infix wildcard searches for TEXT and TAG fields. This enhancement provides greater flexibility in data retrieval and filtering.

t-digest: a new probabilistic data structure for quantile estimation

Redis introduces t-digest, an advanced probabilistic data structure that efficiently estimates quantiles in large datasets or continuous data streams. This is particularly beneficial for analytics and monitoring applications where quantile calculations are required.

Retrieve aggregation results for ongoing time series buckets

A new feature allows users to retrieve the latest, still-open time series buckets during compaction.

Time-weighted average aggregator for time series

Redis now includes a time-weighted average aggregator, improving accuracy in average-over-time calculations. This feature is especially valuable for time series data with irregular sampling intervals.

Gap-filling for time series data

To improve time series analytics, Redis introduces gap-filling capabilities. This feature allows interpolation of missing values or repetition of the last known value for empty time buckets, ensuring continuity in time series analysis.

Improvements

Existing data structures

Redis 6.2 introduces over 25 new commands, fulfilling long-standing community requests. Notably:

  • The long-awaited ZUNION and ZINTER commands now allow direct retrieval of results, unlike ZUNIONSTORE and ZINTERSTORE, which store results in a key.
  • Redis streams enhancements include:
    • Support for exclusive range queries, providing finer control over data retrieval.
    • The ability to filter pending messages based on idle time, improving message management.
    • A new mechanism to automatically claim pending messages from a stream consumer group, transferring ownership of messages that have exceeded their idle timeout to a new consumer without requiring manual acknowledgment.

Component versions

The Redis version 6.2 is built from the following component versions: