How are searches implemented in a Flat File CMS

Question

Flat file CMS's don't use databases. So how are searches implemented? Is searching more or less computationally expensive with this type of setup compared to a database powered search?

Answer 1

The problem with a static site and search together is that one is by definition static, while the other is highly dynamic. So out of the box there is no simple way to make the two live happily together.

Flat file CMS arn't static websites. While parsing files is more costly than parsing databases (usually?), a search functionality can easily be provided by the underlying CMS. Look for plugins that can provide what you want.

However, there is some non trivial solutions that can achieve what you want, depending on your infrastructure and your volumetry and if you site can achieve server side computations or not (grav can, gatsby and hugo can't).

The simplest way to do it is to create an index of all your content in a special file, then load that and do the search client side. You can even use already made package to speedup dev time on this option. (for example: https://www.npmjs.com/package/react-fuzzy-search )

The pro is that it's quite trivial to do. the cons are that the index will get quite big with large side and all the search is done client side (so, maybe a long waiting time for the user if the index is large enough). This solution will also NOT scale well.

Another way to do it is to use a search service (as a SAAS or on your own premises) to externalize the search functionality. Basically this service run a your server, will have a way to index your content (via an API) and search ie (via an API). Just make sure the search API is public and you can query it in realtime from client side.

This solution scales really well because these sort of services are made from the ground up to scale ! However the setup costs are really high, and not worth it if you don't plan to scale to millions of pages.