---
title: Filtering
---

Filters over numeric, boolean, or datetime fields that are indexed as [fast fields](/documentation/indexing/fast_fields) can be "pushed down" into the BM25 index using the `@@@` operator. For instance, the following
two queries produce identical results, but the second query "pushes down" the `rating > 2` filter into the BM25 index.

```sql
-- Without BM25 pushdown
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND rating > 2;

-- With BM25 pushdown
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND rating @@@ '>2';
```

<Note>
  Automatically executing the first type of query as the second query is on the
  roadmap, but not yet supported.
</Note>

## Numeric Filter

The following query returns all rows matching `description:shoes` where `rating` equals `4`.

```sql
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND rating @@@ '4';
```

Comparison operators `>`, `>=`, `<`, and `<=` are supported.

```sql
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND rating @@@ '>=4';
```

## Datetime Filter

RFC 3339 formatted dates wrapped in double quotes can be used to filter over datetime fields.

```sql
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND created_at @@@ '"2023-04-20T16:38:02Z"';
```

## Boolean Filter

Boolean fields can be filtered with `true` or `false`.

```sql
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND in_stock @@@ 'true';
```

## Range Filter

The inclusive `[]` and exclusive `{}` bounds can be used for range filtering.

```sql
-- Numeric
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND rating @@@ '[1 TO 4]';

-- Datetime
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND created_at @@@ '[2020-01-31T00:00:00Z TO 2024-01-31T00:00:00Z]';
```

Range filters can also be applied over text fields. The following query finds all documents where
`description` contains a token that is lexicographically between `book` and `camera` (i.e. would come between
`book` and `camera` in a dictionary).

```sql
SELECT description, rating, category
FROM mock_items
WHERE description @@@ '[book TO camera]';
```

## Set Filter

The `IN` operator finds documents that match at least one element in a set.

```sql
SELECT description, rating, category
FROM mock_items
WHERE description @@@ 'shoes' AND rating @@@ 'IN [2 3 4]';
```