Avoid the Parent/Child Trap: Tips and Tricks for Denormalizing Solr Data

Solr has multiple ways of dealing with parent-child relationships, and each have their own features and drawbacks. Classic Joins have well documented performance issues, especially on larger datasets. Block-Joins require additional indexing work and still increase document count which affects performance to a lesser degree. Full denormalization (having only parents as Solr documents) can be difficult to implement without losing functionality. Recently, we engaged with a client to move them from a Block-Join configuration to a full denormalized solution. Here are some of the tips and tricks to consider if you are looking to make a similar move. Read more →