---
toc_priority: 45
toc_title: s3
---

# S3 表函数 {#s3-table-function}

提供类似于表的接口来 select/insert [Amazon S3](https://aws.amazon.com/s3/)中的文件。这个表函数类似于[hdfs](../../sql-reference/table-functions/hdfs.md)，但提供了 S3 特有的功能。

**语法**

``` sql
s3(path, [aws_access_key_id, aws_secret_access_key,] format, structure, [compression])
```

**参数**

-   `path` — 带有文件路径的 Bucket url。在只读模式下支持以下通配符: `*`, `?`, `{abc,def}` 和 `{N..M}` 其中 `N`, `M` 是数字, `'abc'`, `'def'` 是字符串. 更多信息见[下文](#wildcards-in-path).
-   `format` — 文件的[格式](../../interfaces/formats.md#formats).
-   `structure` — 表的结构. 格式像这样 `'column1_name column1_type, column2_name column2_type, ...'`.
-   `compression` — 压缩类型. 支持的值: `none`, `gzip/gz`, `brotli/br`, `xz/LZMA`, `zstd/zst`. 参数是可选的. 默认情况下，通过文件扩展名自动检测压缩类型.

**返回值**

一个具有指定结构的表，用于读取或写入指定文件中的数据。

**示例**

从 S3 文件`https://storage.yandexcloud.net/my-test-bucket-768/data.csv`中选择表格的前两行:

``` sql
SELECT *
FROM s3('https://storage.yandexcloud.net/my-test-bucket-768/data.csv', 'CSV', 'column1 UInt32, column2 UInt32, column3 UInt32')
LIMIT 2;
```

``` text
┌─column1─┬─column2─┬─column3─┐
│       1 │       2 │       3 │
│       3 │       2 │       1 │
└─────────┴─────────┴─────────┘
```

类似的情况，但来源是`gzip`压缩的文件:

``` sql
SELECT *
FROM s3('https://storage.yandexcloud.net/my-test-bucket-768/data.csv.gz', 'CSV', 'column1 UInt32, column2 UInt32, column3 UInt32', 'gzip')
LIMIT 2;
```

``` text
┌─column1─┬─column2─┬─column3─┐
│       1 │       2 │       3 │
│       3 │       2 │       1 │
└─────────┴─────────┴─────────┘
```

## 用法 {#usage-examples}

假设我们在S3上有几个文件，URI如下:

-   'https://storage.yandexcloud.net/my-test-bucket-768/some_prefix/some_file_1.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/some_prefix/some_file_2.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/some_prefix/some_file_3.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/some_prefix/some_file_4.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/another_prefix/some_file_1.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/another_prefix/some_file_2.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/another_prefix/some_file_3.csv'
-   'https://storage.yandexcloud.net/my-test-bucket-768/another_prefix/some_file_4.csv'

计算以数字1至3结尾的文件的总行数:

``` sql
SELECT count(*)
FROM s3('https://storage.yandexcloud.net/my-test-bucket-768/{some,another}_prefix/some_file_{1..3}.csv', 'CSV', 'name String, value UInt32')
```

``` text
┌─count()─┐
│      18 │
└─────────┘
```

计算这两个目录中所有文件的行的总量:

``` sql
SELECT count(*)
FROM s3('https://storage.yandexcloud.net/my-test-bucket-768/{some,another}_prefix/*', 'CSV', 'name String, value UInt32')
```

``` text
┌─count()─┐
│      24 │
└─────────┘
```

!!! warning "Warning"
    如果文件列表中包含有从零开头的数字范围，请对每个数字分别使用带括号的结构，或者使用`?`。

计算名为 `file-000.csv`, `file-001.csv`, … , `file-999.csv` 文件的总行数:

``` sql
SELECT count(*)
FROM s3('https://storage.yandexcloud.net/my-test-bucket-768/big_prefix/file-{000..999}.csv', 'CSV', 'name String, value UInt32');
```

``` text
┌─count()─┐
│      12 │
└─────────┘
```

插入数据到 `test-data.csv.gz` 文件:

``` sql
INSERT INTO FUNCTION s3('https://storage.yandexcloud.net/my-test-bucket-768/test-data.csv.gz', 'CSV', 'name String, value UInt32', 'gzip')
VALUES ('test-data', 1), ('test-data-2', 2);
```

从已有的表插入数据到 `test-data.csv.gz` 文件:

``` sql
INSERT INTO FUNCTION s3('https://storage.yandexcloud.net/my-test-bucket-768/test-data.csv.gz', 'CSV', 'name String, value UInt32', 'gzip')
SELECT name, value FROM existing_table;
```

**另请参阅**

-   [S3 引擎](../../engines/table-engines/integrations/s3.md)

[原始文章](https://clickhouse.tech/docs/en/sql-reference/table-functions/s3/) <!--hide-->