455: Noisy Animals Kata with Fritz Meissner

Hello, and welcome to another episode of The Bike Shed, a weekly podcast from your friends at Thoughtbot about developing great software. I'm Joelle Kenville. And today, I'm joined by fellow Thoughtbotter, Fritz Meissner. Hello. And together, we're here to share a bit of what we've learned along the way. So, Fritz, what's new in your world? Well, I am looking at optimizing a Postgres query. I have a query that runs a lot of the time. It's fairly fast, but it needs to be faster. What I can see is that it's not using the index that I think it should. I can see the index is there, so now I'm trying to create a new index that Postgres will understand. That's interesting. How can you tell that it's not using the index? With an explain plan. So if you say, explain, analyze in, side of your Postgres, client, it'll spit back some information about what the query is actually doing, like, what it's doing to retrieve data. And I can see that it is using an index, just not the one I wanted to use. Do you think the query planner is correct in thinking that not using this index is the fastest way to work? Or do you think this is the case where you as a developer, like, no. It would be faster if you use this index. Please use this index and don't do what you actually are hoping to do. Yeah. I'm I'm pretty sure that it's wrong and I'm right. But the only way to really know is to, try. So we'll create the index and then see what happens afterwards. It is no great loss if it doesn't fix the problem or at least it'll be I'll be sad that it isn't solved, but we'll try something else. How often do you run into situations where you think the database query planner is wrong and you're right? This is probably the first time in my career that I've thought that. Okay. Yeah. So I feel like I always tend to blame myself in those situations. I'm like, well, if the query plan is choosing to not use my index, clearly, I am indexing on the wrong thing here, or the dataset is too small or something like that. I need to refine my approach. Yeah. I guess it is pretty bold to assume that all of the knowledge that's baked into Postgres is, wrong, and my grasp of it is better. What's new in your world? Oh, I've been recently fighting with some tricky unit tests. Yeah. This is an RSpec test, and I'm parsing data from a third party API that comes in as JSON, and it gets turned into a pretty complex series of nested active model objects. Yeah. And my assertion on this parser is that it will return I give it some JSON. I parse it, and then I expect the response to have certain attributes, using our …