mirror of
https://github.com/ash-project/ash_archival.git
synced 2024-09-21 05:52:56 +12:00
31 lines
1.1 KiB
Markdown
31 lines
1.1 KiB
Markdown
# Upserts & Identities
|
|
|
|
Its important to consider identities when using AshArchival _without_ a `base_filter` set up.
|
|
|
|
If you are using a `base_filter`, then all identities implicitly include that `base_filter` in their
|
|
`where` (handled by the data layer).
|
|
|
|
Take the following identities, for example:
|
|
|
|
```elixir
|
|
identities do
|
|
identity :unique_email, [:email], where: expr(is_nil(archived_at))
|
|
# and
|
|
identity :unique_email, [:email]
|
|
end
|
|
```
|
|
|
|
## With `is_nil(archived_at)`
|
|
|
|
Using this identity allows multiple archived records with the same email, but only one _non-archived_ record per email.
|
|
It enables reuse of archived email addresses for new active records, maintaining data integrity by preventing duplicate
|
|
active records while preserving archived data.
|
|
|
|
When you upsert a record using this identity, it will only consider active records.
|
|
|
|
## Without `is_nil(archived_at)`
|
|
|
|
This identity configuration enforces strict email uniqueness across all records. Once an email is used, it can't be used
|
|
again, even after that record is archived.
|
|
|
|
When you upsert a record using this identity, it will consider all records.
|