0

I would like to improve search functionality on our magento website: AND instead of OR in the default search field. Search suggestions and spelling corrections would be great.

Searching for this, I found the following modules: Elastic Search (http://www.magentocommerce.com/magento-connect/elasticsearch.html) and Sphinx Search Pro.

Elastic search is "search in the cloud", but I have no idea what that means here. Our magento installation is installed on AWS EC2, so in the cloud, but I don't see how that makes a difference. Search is performed on the database, or does it work otherwise, from cache or somehow?

What is the best module to use? Any other tips?

Improve this question
1
  • 1
    Magento internal search is <50% accurate, Sphinx is 60-80% accurate, Solr is 98% accurate (we use Solr). ElasticSearch is a new one but there are many cloud search engines, Cloudera, etc., normally costing high $100s/mth so only for $multi-million companies, search should be around 0.5% of revenue so if there is a yearly cost you need to build this in.
    – user2935
    Commented Mar 31, 2014 at 16:39

4 Answers 4

Reset to default
1

Sphinx Search Pro (or more extended version Sphinx Search Ultimate) very powerful solutions, which return highlevel relevant results.

For use these modules by default you no need install any additional software to server (+ these extensions not cloud servers, etc). All search performs on your server.

Additionally these plugins support search by multiple content types (product, categories, blogs, articles etc) http://ssu.demo.mirasvit.com/catalogsearch/result/?cat=0&q=shoes

Improve this answer
1
  • It took a while, but last week I installed this module. Installation was easy, and it works like expected. The dropdown search box is a real nice plus. So while I found the solution before you answered my question, it's still good enough to accept this as the answer that works best for me. The other suggestions seem too complex, especially SOLR.
    – SPRBRN
    Commented Jun 23, 2014 at 9:33
7

I've used both Elastic Search and SOLR on a few projects. Solr is very accurate but a bit more tricky to configure and the Magento SOLR Bridge extension had some issues with multistore setup.

After SOLR I tried Elastic Search which was quite easy to configure and was up and running in an hour. This included installing Elastic Search on the server itself and getting it up and running on Magento.

Both platforms can be used in the cloud but also installed on your local Linux machine. How ever both require quite some resources so make sure your machine has enough.

On which one is better you might want to read this Stackoverflow answer and browse through this page.

Ultimately any extension you install for search in Magento will be an improvement because the default MySQL based search engine is a disaster. Personally I'm a fan of Elastic Search but again, read the links and decide for yourself.

Improve this answer
4
1

Blast Lucene uses the Lucene search already built into the Zend Framework, which Magento is partially built on. It does an excellent job of indexing Magento product and allows adding weighting to any attributes that are included in "Quick Search".

If you add a "supplemental search terms" textarea attribute, you can have full control on returning search results for sku variants, part number cross reference, industry jargon and alternate names used for product.

Highly relevant returns and finally, the sort by relevance also works.

Improve this answer
1
1

Another choice for Sphinx is sphinxsearch by fheyer. It is free, and can be found here : https://github.com/fheyer/sphinxsearch

I have used this on several sites with great results. I haven't installed it on anything later than 1.8, so ymmv, but I would think it works or could be made to work with newer instances.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.