---
title: Overview
---

<Note>
Aggregate pushdown is currently a beta feature.
</Note>

## Basic Usage

ParadeDB can accelerate certain styles of aggregates by pushing them down to the BM25 index. This feature is disabled by default and can be enabled by running

```sql
SET paradedb.enable_aggregate_custom_scan TO on;
```

With this setting enabled, queries like the following can be pushed down:

```sql
SELECT AVG(rating) FROM mock_items
WHERE id @@@ paradedb.all();
```

To verify that a query is pushed down, look for `ParadeDB Aggregate Scan` in the `EXPLAIN` output:

```sql
EXPLAIN SELECT AVG(rating) FROM mock_items
WHERE id @@@ paradedb.all();
```

```csv Expected Response
                                       QUERY PLAN
----------------------------------------------------------------------------------------
 Custom Scan (ParadeDB Aggregate Scan) on mock_items  (cost=0.00..0.00 rows=0 width=32)
   Index: search_idx
   Tantivy Query: {"with_index":{"query":"all"}}
   Aggregate Definition: {"0":{"avg":{"field":"rating"}}}
(4 rows)
```

We are actively working on expanding the types of aggregates that can be pushed down. The following aggregates are currently supported:
[count](/documentation/aggregates/metrics/count), [sum](/documentation/aggregates/metrics/sum), [min/max](/documentation/aggregates/metrics/minmax),
[average](/documentation/aggregates/metrics/average), and [terms](/documentation/aggregates/bucket/terms).