---
title: Common Aggregates
---

<Note>
  Please read the [overview](/documentation/aggregates/overview) for important information on how to accelerate aggregate queries.
</Note>

## Count

`COUNT` tallies the number of matching rows/documents from the table.

```sql
SELECT COUNT(*) FROM mock_items
WHERE description @@@ 'shoes';
```

<Info>
Because `COUNT(*)` does not specify any fields, you do not need to worry about passing a fast field to `COUNT` if `*` is used.
</Info>

## Average

`AVG` calculates the mean of the specified numeric field values across all documents.

```sql
SELECT AVG(rating) FROM mock_items
WHERE description @@@ 'shoes';
```

## Sum

`SUM` computes the total sum of the specified numeric field values across all documents.

```sql
SELECT SUM(rating) FROM mock_items
WHERE description @@@ 'shoes';
```

## Min

`MIN` finds the smallest value for the specified numeric field across all documents.

```sql
SELECT MIN(rating) FROM mock_items
WHERE description @@@ 'shoes';
```

## Max

`MAX` finds the largest value for the specified numeric field across all documents.

```sql
SELECT MAX(rating) FROM mock_items
WHERE description @@@ 'shoes';
```

## Cardinality

`COUNT(DISTINCT)` returns the number of unique values in a field. This is also known as a cardinality aggregation.

```sql
SELECT COUNT(DISTINCT rating) FROM mock_items
WHERE description @@@ 'shoes';
```

## Histogram

The following query creates buckets of size `1` for the `rating` field. Each field value is rounded down to its bucket.

```sql
SELECT FLOOR(rating / 1) * 1 AS bucket, COUNT(*) AS count
FROM mock_items
GROUP BY bucket
ORDER BY bucket;
```

<Accordion title="Expected Response">
```csv
 bucket | count
--------+-------
      1 |     1
      2 |     3
      3 |     9
      4 |    16
      5 |    12
(5 rows)
```
</Accordion>

This query increases the bucket size to `2`.

```sql
SELECT FLOOR(rating / 2) * 2 AS bucket, COUNT(*) AS count
FROM mock_items
GROUP BY bucket
ORDER BY bucket;
```

<Accordion title="Expected Response">
```csv
 bucket | count
--------+-------
      0 |     1
      2 |    12
      4 |    28
(3 rows)
```
</Accordion>

## Other Aggregates

While the above list of aggregates is non-exhaustive, the answer to any kind of aggregate in ParadeDB is to just write SQL. Any aggregate that can be expressed
in SQL can be written in ParadeDB.