Which CouchDB application would be most useful / effective?

Question

I’ve already fumbled with the Ordnance Survey Code-Point Open Overview - a UK Postal Code / Coordinates dataset. Since Couch.io offered a free copy of CouchDB, I thought I would put my geodata in one of them, learning a little about CouchDB in the process.

The idea was that since CouchDB should handle large datasets well (zip code is about 1.7 million records) and works with REST / JSON, it will be well connected with client-side jQuery for use with Google Apps for cards.

My initial goal was to simply make an AJAX call with a zip code as a parameter, returning a single JSON object with lat / lon parameters that I could use in my script (showing a token for this zip code).

I did it successfully, but based on the database relational database, it was much more complicated than I thought; since I read more about CouchDB and play a little with it, I get the impression that this would not be the right tool for this job if I used it for a real project.

Do I think that dynamic queries are a bit of a weakness for CouchDB? Is it more focused on returning large views that often do not change from large data sets? What could be some examples of the “good” and “bad” uses of CouchDB in terms of its strengths?

Answer 1

I am the main guy at Couchio. Glad you enjoy CouchDB.

I feel that basically relational databases are better suited for ever-changing one-time queries from large datasets. For all this time, brute force is still required. Neither SQL nor NoSQL is a silver bullet. However, in a broad sense, NoSQL databases are better if you already know what questions you will ask. In other words, it is not a question of how much data is changing, but how many changes are in the request.

This is a theory. For your specific project, is CouchDB suitable? I feel that there is nothing wrong with doing a lot of indexes on a basic dataset. The advantage of queries is only by index; queries occur very quickly. CouchDB, in particular, only needs to re-index new data, even for queries such as averages or XOR checksums.

So, even if you have hundreds of different types of queries that you could fulfill, if you already know what those queries are, just write them. However, if you never stop creating new queries, CouchDB will not be easy to keep up.