Multiple queries when adding tags

[mister-rao]

Inspecting with django-toolbar I found that taggit makes n*2 queries for adding n number of tags.

Is there any way to optimize this?

[rtpg]

Could you post what kind of tags are shown to be made? It might be a bit endemic to how Django models work by default

[mister-rao]

This is my request body:

{
  "name": "New Collection",
  "type": "movie",
  "description": "This is a new movie collection",
  "tags": [
    "movie", "new", "awesome", "binge", "adventure", "thriller", "epic", "fantasy", "dark"
  ]
}

this is how I add tags:

collection.tags.add([ "movie", "new", "awesome", "binge", "adventure", "thriller", "epic", "fantasy", "dark"])

These are the queries:

[rtpg]

OK so if you know what you are doing and don't use save hooks or the like and really know what you are doing , you can use bulk_create and bulk_update (Django ORM methods) to do tag creation manually and create 10 tags in a single query (or at least less queries).

The problem with django-taggit doing that is that would mean that tags would not be able to have save methods. This is basically the performance/usability tradeoff present in all Django systems.

django Taggit's "create N tags" is really just a shortcut for "create 1 tag, but do it for n objects". We're doing lookups to find existing tags, and then calling .save on each tag. Kinda hard to reduce that. And if we call bulk_create/bulk_update, we'll break code that relies on .save being called, or post/pre-save signals on tags. That's why it's more or less up to users to do this themselves.

Having said that, I think this could be a good FAQ entry. I've opted for bulk_X as a solution to this problem as well, so I understand the need in the abstract.

(And as usual: this is all based on the assumption that this is a performance issue on your project to begin with. Measuring the impact on production is so important)