# XREADGROUP

> Read as consumer group.

Use `XREADGROUP` to read entries from a stream as a member of a consumer group, so that each entry goes to one consumer and Redis keeps track of what has not been acknowledged.

After `STREAMS`, all keys come first and then one ID per key. The special ID `>` delivers entries that have never been delivered to the group and records them as pending for this consumer. Any other ID re-reads that consumer's own pending entries instead, starting after the given ID, which is how a consumer resumes after a restart: read from `0` first to finish old work, then switch to `>`.

The consumer is created on first use. `COUNT` limits the batch size and `BLOCK <milliseconds>` waits for new entries rather than returning empty, with `0` waiting indefinitely. `NOACK` skips the pending entries list entirely, trading the delivery guarantee for speed.

Entries stay pending until acknowledged with [`XACK`](/redis/commands/streams/xack). That is what makes recovery possible: work left behind by a crashed consumer is visible in [`XPENDING`](/redis/commands/streams/xpending) and can be taken over with [`XAUTOCLAIM`](/redis/commands/streams/xautoclaim).

## Syntax

```redis
XREADGROUP GROUP <group> <consumer>
  [COUNT <count>]
  [BLOCK <milliseconds>]
  [NOACK]
  STREAMS <key> [<key> ...] <ID> [<ID> ...]
```

## Arguments

| Argument | Required | Repeatable | Description |
| --- | --- | --- | --- |
| `GROUP <group> <consumer>` | Yes | No | Consumer group to read from, and the consumer within it that claims the entries. |
| `COUNT <count>` | No | No | Maximum number of entries to return per stream. |
| `BLOCK <milliseconds>` | No | No | Milliseconds to block waiting for new entries; `0` blocks indefinitely. |
| `NOACK` | No | No | Do not add delivered entries to the pending list. |
| `STREAMS <key> [<key> ...] <ID> [<ID> ...]` | Yes | No | Streams to read. List every key first, then one ID per key in the same order. |

## Important points

- A blocking form holds the request until data arrives or its timeout expires. Set the client/network timeout longer than the command timeout.
- After `STREAMS`, provide all keys first and then exactly one ID for each key, in the same order.
- `COUNT` and `BLOCK` each require a value. Giving either without one returns a wrong number of arguments error.

## Response

The reply reports the result of the operation. Error replies have the same shape in RESP2 and RESP3 and are surfaced as exceptions by the SDKs below.

| Protocol | Reply |
| --- | --- |
| RESP2 | Null bulk string or null array or Flat array of alternating keys and values |
| RESP3 | Null or Map |

<Note>
  Client libraries often decode bulk strings, maps, sets, and numeric strings into language-native values. The table describes the Redis wire reply.
</Note>

## Examples

TCP examples use the TLS `REDIS_URL` from the Upstash console. REST examples use `UPSTASH_REDIS_REST_URL` and `UPSTASH_REDIS_REST_TOKEN`.

<AccordionGroup>

<Accordion title="Redis CLI" icon="terminal">

```bash
XREADGROUP GROUP workers worker-1 STREAMS my-key 0-0
```

</Accordion>

<Accordion title="@upstash/redis" icon="node-js" iconType="brands">

```ts
import { Redis } from "@upstash/redis";

const redis = Redis.fromEnv();

const result = await redis.xreadgroup("mygroup", "consumer1", "mystream", ">");
```

</Accordion>

<Accordion title="upstash_redis" icon="python" iconType="brands">

```python
from upstash_redis import Redis

redis = Redis.from_env()
result = redis.xreadgroup("workers", "worker-1", {"my-key": "0-0"})
print(result)
```

</Accordion>

<Accordion title="ioredis" icon="node-js" iconType="brands">

```ts
import Redis from "ioredis";

const redis = new Redis(process.env.REDIS_URL!);
const result = await redis.xreadgroup("GROUP", "workers", "worker-1", "STREAMS", "my-key", "0-0");
console.log(result);
```

</Accordion>

<Accordion title="node-redis" icon="node-js" iconType="brands">

```ts
import { createClient } from "redis";

const client = await createClient({ url: process.env.REDIS_URL })
  .on("error", console.error)
  .connect();
const result = await client.xReadGroup("workers", "worker-1", { key: "my-key", id: "0-0" });
console.log(result);
```

</Accordion>

<Accordion title="redis-py" icon="python" iconType="brands">

```python
import os
import redis

client = redis.from_url(os.environ["REDIS_URL"])
result = client.xreadgroup("workers", "worker-1", {"my-key": "0-0"})
print(result)
```

</Accordion>

<Accordion title="go-redis" icon="golang" iconType="brands">

```go
package main

import (
    "context"
    "fmt"
    "os"

    "github.com/redis/go-redis/v9"
)

func main() {
    opts, err := redis.ParseURL(os.Getenv("REDIS_URL"))
    if err != nil {
        panic(err)
    }
    client := redis.NewClient(opts)
    result, err := client.XReadGroup(context.Background(), &redis.XReadGroupArgs{Group: "workers", Consumer: "worker-1", Streams: []string{"my-key", "0-0"}}).Result()
    if err != nil {
        panic(err)
    }
    fmt.Println(result)
}
```

</Accordion>

<Accordion title="jedis" icon="java" iconType="brands">

```java
import java.net.URI;

import redis.clients.jedis.Jedis;

try (Jedis jedis = new Jedis(new URI(System.getenv("REDIS_URL")))) {
  Object result = jedis.xreadGroup("workers", "worker-1", redis.clients.jedis.params.XReadGroupParams.xReadGroupParams(), java.util.Map.of("my-key", new redis.clients.jedis.StreamEntryID("0-0")));
  System.out.println(result);
}
```

</Accordion>

<Accordion title="redis-rs" icon="rust" iconType="brands">

```rust
fn main() -> redis::RedisResult<()> {
    let url = std::env::var("REDIS_URL").expect("REDIS_URL is not set");
    let client = redis::Client::open(url)?;
    let mut connection = client.get_connection()?;

    let mut command = redis::cmd("XREADGROUP");
    command.arg("GROUP");
    command.arg("workers");
    command.arg("worker-1");
    command.arg("STREAMS");
    command.arg("my-key");
    command.arg("0-0");
    let result: redis::Value = command.query(&mut connection)?;
    println!("{result:?}");
    Ok(())
}
```

</Accordion>

</AccordionGroup>