Update nested struct with Ecto

Two options. Either just do inserts directly to Files that include asset_id, or if you will always have the entire collection of files loaded, you can use nested changesets. So in your Asset changeset, you will have a function: cast_assoc(:files, required: false, with: &File.changeset/2)

When you update Asset though, you need to include at least the id for all existing files, as well as the fields you want to set for any new files. The former option is the only safe one if concurrent updates are possible; by default you will get an exception if you leave an ID out of your new payload.

Ok, now i have this:

params = %{
  files: [
    %{filename: "name1"},
    %{filename: "name2"} 
  ]
}

asset
|> Repo.preload(:files)
|> Ecto.Changeset.cast(params, []) 
|> Ecto.Changeset.cast_assoc(:files) # works o only at first time
|> Repo.update

So first time it works well, but second time it raises exception:
** (RuntimeError) you are attempting to change relation :files of Asset but the :on_replace option of this relation is set to :raise. By default it is not possible to replace or delete embeds and associations during cast. Therefore Ecto requires all existing data to be given on update. Failing to do so results in this error message. If you want to replace data or automatically delete any data not sent to cast, please set the appropriate :on_replace option when defining the relation. The docs for Ecto.Changeset covers the supported options in the "Related data" section. However, if you don't want to allow data to be replaced or deleted, only updated, make sure that:

• If you are attempting to update an existing entry, you are including the entry primary key (ID) in the data.
• If you have a relationship with many children, at least the same N children must be given on update.

So i I thought i need to set on_replace: :update, but update option is available only for has_one association.

Yes this was what I was referring to. That means there are existing associations you have not included in your payload. After the first execution there are two files that were assigned IDs. You need to include those in your next params:

params = %{
  files: [
    %{id: file1.id},
    %{id: file2.id},
    %{filename: "name3"},
    %{filename: "name4"} 
  ]
}

yes it really works. Thank u.

Do these updates happen atomically? Or if something goes wrong mid changes you end up with half of the updates?

I just tested it and it wraped the insert in a transaction. if one of subsequent operations fails, it rolls back the previous changes.