> 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/querying/the-collection-object/filter-the-query-1/filter-the-query.md).

# Filter the query – The Expression Engine

Collections represent SQL `SELECT` queries. Lustra's expression engine lets you write SQL conditions with Crystal syntax while still generating PostgreSQL SQL.

## Tuple and Hash Conditions

For simple equality checks, use named arguments, a `NamedTuple`, or a hash:

```crystal
User.query.where(first_name: "Richard")
User.query.where({first_name: "Richard", active: true})
```

Tuple/hash conditions also handle arrays, ranges, nil, and subqueries:

```crystal
User.query.where(id: [1, 2, 3])
User.query.where(created_at: 1.week.ago..Time.utc)
User.query.where(deleted_at: nil)

subquery = Post.query.select("user_id").where(published: true)
User.query.where(id: subquery)
```

## Template Conditions

Use templates when a small SQL fragment is clearer. Placeholder values are escaped:

```crystal
User.query.where("first_name = ? OR last_name = ?", "Richard", "Richard")
User.query.where("first_name = :name OR last_name = :name", name: "Richard")
```

Avoid interpolating user input directly into the SQL string.

## Expression Blocks

Use an expression block for richer conditions:

```crystal
User.query.where { (first_name == "Richard") | (last_name == "Richard") }
User.query.where { created_at.in?(1.week.ago..Time.utc) }
User.query.where { posts_count.between(10, 20) }
User.query.where { email =~ /@example\.com$/i }
```

Common operators:

| Crystal                          | SQL meaning                                                          |
| -------------------------------- | -------------------------------------------------------------------- |
| `==` / `!=`                      | equality / inequality, including `IS NULL` and `IS NOT NULL` for nil |
| `<`, `<=`, `>`, `>=`             | comparison                                                           |
| `=~`, `!~`                       | PostgreSQL regex match / non-match                                   |
| `&`, `\|`                        | `AND`, `OR`                                                          |
| `~condition` or `not(condition)` | `NOT`                                                                |
| `in?(array)`                     | `IN (...)`                                                           |
| `in?(range)`                     | range comparison                                                     |
| `in?(subquery)`                  | `IN (SELECT ...)`                                                    |
| `between(a, b)`                  | `BETWEEN a AND b`                                                    |

Because Crystal does not let libraries redefine `&&` and `||`, use `&` and `|` for SQL `AND` and `OR`.

Always parenthesize each side of `&` and `|`:

```crystal
User.query.where { (active == true) & (posts_count > 0) }
User.query.where { (role == "admin") | (role == "owner") }
```

## Local Variables and Column Names

The expression engine uses `method_missing` to turn unknown names into SQL variables. If a local variable has the same name as a column, Crystal resolves the local variable first:

```crystal
def find_user(id)
  User.query.where { id == id } # wrong: both sides are the local variable
end
```

Use `var` to force a SQL column:

```crystal
def find_user(id)
  User.query.where { var("id") == id }
end
```

`var` quotes each identifier part:

```crystal
User.query.where { var("users", "id") == 1 }
# "users"."id" = 1
```

## Raw SQL

Use `raw` for trusted SQL fragments that cannot be expressed through the normal API:

```crystal
User.query.where { raw("LOWER(email)") == "admin@example.com" }
User.query.where { raw("COUNT(*)") > 5 }
```

`raw` pastes SQL into the query. Prefer placeholders for values:

```crystal
User.query.where { raw("LOWER(email) = ?", email.downcase) }
User.query.where { raw("LOWER(email) = :email", email: email.downcase) }
```

{% hint style="danger" %}
Do not build `raw` SQL by interpolating untrusted input. Use placeholders or regular expression/tuple conditions instead.
{% endhint %}

## Custom Operators

Use `op` for PostgreSQL operators that do not map cleanly to Crystal operators:

```crystal
Event.query.where { op(payload, raw("'source'"), "?") }
```

For JSONB-specific querying, prefer Lustra's JSONB helpers where possible.

## Negation and OR

`where.not` wraps a condition in `NOT`:

```crystal
User.query.where.not(active: true)
User.query.where.not { deleted_at == nil }
```

`or` combines the current condition with a new condition:

```crystal
User.query.where(active: true).or(role: "admin")
User.query.where { posts_count > 10 }.or { role == "owner" }
```


---

# 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/querying/the-collection-object/filter-the-query-1/filter-the-query.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.