---
title: AggregateFunction Data Type reference
meta:
    description: Documentation for the AggregateFunction data type.
headingMaxLevels: 2
---

# AggregateFunction

{% callout type="info" %}
This data type isn't supported at ingest. It is only supported at query time and to create Copy Data Sources or Materialized View Data Sources.
{% /callout %}

The `AggregateFunction` data type stores the intermediate state of an aggregate function. This allows for partial aggregation results to be stored and later combined to produce a final aggregate value. It is commonly used in Materialized Views or Copy Data Sources to pre-aggregate data.

To produce a value of this type, you typically use an aggregate function with the `-State` suffix. To retrieve the final aggregated result from one or more `AggregateFunction` states, you use the corresponding aggregate function with the `-Merge` suffix.

### Syntax

```sql
AggregateFunction(name, types_of_arguments...)
```

### Parameters

-   `name`: The name of the aggregate function whose state is being stored (e.g., `uniq`, `sum`, `quantiles`). If the aggregate function itself takes parameters (like `quantiles(0.5)`), these should be included here.
-   `types_of_arguments`: The data types of the arguments that the aggregate function expects (e.g., `UInt64`, `String`).

### Example

This example demonstrates how to first generate an `AggregateFunction` state using `uniqState` and then combine and finalize it using `uniqMerge`.

```sql
SELECT
    uniqMerge(state_col) AS final_unique_count
FROM
(
    SELECT
        uniqState(number) AS state_col
    FROM numbers(5)
    GROUP BY number % 2 -- Creates two intermediate states: one for even numbers, one for odd
) AS subquery;
```

Result:

```result
┌─final_unique_count─┐
│                  5 │
└────────────────────┘
```