> For the complete documentation index, see [llms.txt](https://crystal-garage.gitbook.io/lustra-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://crystal-garage.gitbook.io/lustra-docs/model/lifecycle/persistence.md).

# Persistence

## Save

`save` inserts or updates a model:

```crystal
user = User.new({email: "user@example.com"})

if user.save
  puts user.id
else
  puts user.print_errors
end
```

`save` returns `true` when the record was persisted and `false` when validations fail.

`save!` raises `Lustra::Model::InvalidError` when validations fail and returns the model on success:

```crystal
user = User.new({email: "user@example.com"})
user.save!
```

When the model is new, Lustra runs an `INSERT`. When the model is already persisted and has changed columns, Lustra runs an `UPDATE`.

## Update

`update` assigns attributes and calls `save`:

```crystal
user.update(email: "new@example.com")
```

`update!` assigns attributes and calls `save!`:

```crystal
user.update!(email: "new@example.com")
```

## Idempotent Sync

Use `find_or_build` when synchronizing records from an external source. It returns the existing record when one matches the lookup fields, or builds a new unsaved model with those fields assigned.

```crystal
user = User.query.find_or_build(provider: "github", provider_id: github_user.id)

user.login = github_user.login
user.name = github_user.name
user.avatar_url = github_user.avatar_url
user.synced_at = Time.utc
user.save!
```

This lets the sync code assign the latest external values and save once.

Use `find_or_create` when the initial values are enough to create the record immediately:

```crystal
tag = Tag.query.find_or_create(name: "orm")
```

`find_or_create` is also useful for join records:

```crystal
RepositoryFork.query.find_or_create(
  parent_id: parent_repository.id,
  fork_id: repository.id
)
```

Prefer `find_or_build` when the record needs more fields assigned before saving. Prefer `find_or_create` when the lookup fields are enough for a valid record.

## Reload

`reload` fetches the current database row by primary key, replaces the model values, clears cached association data, and marks the model as persisted:

```crystal
user.reload
```

## Delete and Destroy

`delete` removes the row directly and does not run callbacks:

```crystal
user.delete
```

`destroy` removes the row and runs `:destroy` callbacks:

```crystal
user.destroy
```

Use `destroy` when application cleanup belongs in callbacks. Use `delete` only when you intentionally want a direct delete.

## Import

`Model.import` inserts many new models in one SQL query:

```crystal
users = [
  User.new({email: "one@example.com"}),
  User.new({email: "two@example.com"}),
]

saved_users = User.import(users)
```

Each imported model must be new. Lustra validates each model and runs save/create callbacks around the import.

## Read-Only Models

Set `self.read_only = true` for models that should not be saved, such as database views or system catalog models:

```crystal
class RepositoryStats
  include Lustra::Model

  self.table = "repository_stats"
  self.read_only = true

  column repository_id : Int64, primary: true
  column forks_count : Int64
end
```

`save` returns `false` for read-only models. `save!` raises `Lustra::Model::ReadOnlyError` with a message explaining that the model is read-only.

## Lifecycle-Bypassing Methods

Some methods intentionally update the database directly. They do not run validations, callbacks, or automatic timestamp updates unless stated otherwise.

| Method                      | Behavior                                                                                   |
| --------------------------- | ------------------------------------------------------------------------------------------ |
| `delete`                    | Deletes one persisted model without callbacks.                                             |
| `Model.insert`              | Inserts one row directly without building a model.                                         |
| `Model.insert_all`          | Inserts many rows directly in one SQL statement.                                           |
| `Model.upsert`              | Inserts or updates one row through PostgreSQL `ON CONFLICT`.                               |
| `Model.upsert_all`          | Inserts or updates many rows through PostgreSQL `ON CONFLICT`.                             |
| `update_column`             | Updates one column directly.                                                               |
| `update_columns`            | Updates multiple columns directly.                                                         |
| `increment!` / `decrement!` | Atomically updates a numeric column in the database.                                       |
| `touch`                     | Updates timestamp columns through `update_columns`; it bypasses validations and callbacks. |
| `Collection#delete_all`     | Bulk deletes matching rows without loading models.                                         |
| `Collection#update_all`     | Bulk updates matching rows without loading models.                                         |

Use lifecycle-bypassing methods for counters, flags, maintenance tasks, and bulk operations where skipping callbacks is intentional.

Use `destroy`, `save`, `save!`, `update`, or `update!` when validations and callbacks matter.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://crystal-garage.gitbook.io/lustra-docs/model/lifecycle/persistence.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.