# Scanning

> For the complete documentation index, see [llms.txt](https://effectkit.dev/llms.txt).

Walk keys, hashes, sets and sorted sets as lazy Streams over cursors.

Source: https://effectkit.dev/docs/redis/scanning

Redis scans return a few entries and a cursor. The scan methods on `redis` hide the cursor and give you a `Stream` that asks for the next page when you pull. Take only what you need and the rest is never fetched.

**Example** (Collecting the First 20 Matching Keys)

```ts twoslash
import { Effect, Stream } from "effect";

import { Redis } from "@effect-kit/redis";

const program = Effect.gen(function* () {
  const redis = yield* Redis;

  const keys = yield* redis
    .scanKeys({ match: "user:*", count: 100 })
    .pipe(Stream.take(20), Stream.runCollect);
  //    ^?

  return keys.map((key) => new TextDecoder().decode(key));
});
```

`match` is a glob pattern and `count` is a hint for the page size. `scanKeys` also takes `type` to filter by key type. Keys are `Uint8Array`, so decode them before use.

<Callout type="info" title="Scans can repeat keys">
  A scan is not a snapshot. A key can appear more than once, and keys added or removed during the
  scan can be missed or seen. Deduplicate if repeats matter.
</Callout>

## Scanning one collection [#scanning-one-collection]

Each collection type has its own stream.

| Method                                | Element type                                                             |
| ------------------------------------- | ------------------------------------------------------------------------ |
| `redis.scanKeys(options?)`            | `Uint8Array`                                                             |
| `redis.scanHash(key, options?)`       | `readonly [field, value]` as `Uint8Array` pairs                          |
| `redis.scanHashFields(key, options?)` | `Uint8Array` field names                                                 |
| `redis.scanSet(key, options?)`        | `Uint8Array` members                                                     |
| `redis.scanSortedSet(key, options?)`  | `readonly [member, score]`, member as `Uint8Array` and score as `string` |

**Example** (Streaming a Sorted Set)

```ts twoslash
import { Effect, Stream } from "effect";

import { Redis } from "@effect-kit/redis";

const program = Effect.gen(function* () {
  const redis = yield* Redis;

  yield* redis.zadd("leaderboard", [
    [1.5, "ada"],
    [2, "lin"],
  ]);

  const entries = yield* Stream.runCollect(redis.scanSortedSet("leaderboard"));
  //    ^?

  return entries;
});
```

Scores come back as strings.

## Paging by hand [#paging-by-hand]

`redis.scan`, `redis.hscan`, `redis.sscan` and `redis.zscan` run one command and return a `ScanPage` with a `cursor` string and the `entries` for that page. The cursor is an opaque decimal string. Use these when you must store the cursor between requests.

On a cluster, `scanKeys` walks the primary nodes one after another. See [scanning a cluster](/docs/redis/cluster/scanning).