When do you use assoc function vs just setting the assoc_id field?

Chat / Discussions
1

Does anyone have any hard and fast rules for when they use the assoc functions vs the assoc_id field on the schema?

Here’s a hypothetical scenario:

I’ve got a Post schema which has a belongs_to of type User. Since I’m never going to be creating a User at the same time as a Post, is there any benefit to using the *_assoc functions, or should I just set the user_id explicitly in the changeset?

Thoughts?

1 Like
2

I think it really depends on whether you have a requirement to save things at the same time (by working with the full association) or not.

I find it more helpful to represent dependencies for relationships that need additional information. This is a silly example, but imagine you have an Item, Recipe, and RecipeItem schema, where RecipeItem contains both the FKs to Item and Recipe but also an amount and units column. Then let’s say we want to store/ dump a JSON full of recipes:

{
  "name": "Mashed Potatoes",
  "recipe_items": [
    {"name": "Potato", "amount": 2}
    {"name": "Milk", "amount": 1, "units": "spoons"}
  ]
}

In this scenario, it’s easier to visualize dependencies and work with the whole thing. if you are not doing that, I don’t really see any advantages, only downsides - like exposing information that you can accidentally change.

Now, If you control the information and don’t need to cast the association, you might as well just use put_assoc or put_change to set either the whole association or the FK (the golden rule is that you cast stuff that is external to the app).