Solr Search for Drupal 8: Lessons Learned
Created in 2004, and donated to the Apache Software Foundation in 2006 for open source publishing of the code, Solr Search has emerged into what many now consider to be the gold standard of enterprise search platforms.
When developing Drupal 8 websites, this highly popular platform built on the Apache Java Lucerne library, is often proven to be the ideal solution and perfect fit for a website’s specific objectives and information architecture.
But not always.
Among the considerations when integrating Solr Search into Drupal 8 sites:
- Solr Search usually requires the set up of a separate server. Drupal search, on the other hand can be set up with the installation of a module and the selection of a few settings. While Solr tends to require considerably more development work for sites that have a wide range of requirements, basic integration can be accomplished with relatively little effort if the Solr server is available.
- Solr often represents the optimal solution for achieving great performance large sites with complex, highly updated content. For less complex websites, Drupal search and caching can also provide good performance.
- Apache Solr comes with a set of great features (listed below), which may or may not be needed. These features will always need to be reviewed against the needs of the project to determine whether they will be useful.
Among Solr’s key achievements: the perception that it’s the best. It’s not uncommon that clients specifically request Solr as the search platform for their site. Our approach is to first seek to understand the big picture, dig deep into requirements, explore available options, and move forward with recommendations.
This exercise is incorporated into the Promet Source architecture workshops. We also create a detailed matrix of search requirements, identifying an entire set of search specifications and then ask the client prioritize them in order to inform the decision-making process.
Prioritized search requirements may include features such as:
- Spell check
- Keyword highlighter
- Multiple character wildcard
- Single character wildcard
- Term proximity
- Exact phrase boost
Range of Objectives
Clearly, one product owner’s idea of what constitutes ideal search capabilities differs from another’s. There is not one gold standard or ideal enterprise search platform, and as such, before discussing the functionality and relative merits of search platform options, we do the work upfront to understand the full scope of search criteria. Until this information has been gathered, one cannot effectively develop recommendations and move forward with decisions based on the search platform that aligns most closely with specific requirements.
Interested in looking beneath the surface concerning the search platform that meets your specific objectives -- or any other aspect of Drupal design and development -- contact us today.
Pamela Ross and Luc Bezier contributed to this post.