S3

Name

S3

description

S3 table-valued-function(tvf), allows users to read and access file contents on S3-compatible object storage, just like accessing relational table. Currently supports csv/csv_with_names/csv_with_names_and_types/json/parquet/orc file format.

grammer

  1. s3(
  2. "uri" = "..",
  3. "access_key" = "...",
  4. "secret_key" = "...",
  5. "format" = "csv",
  6. "keyn" = "valuen",
  7. ...
  8. );

parameter description

Each parameter in S3 tvf is a pair of "key"="value".

Related parameters for accessing S3:

  • uri: (required) The S3 tvf will decide whether to use the path style access method according to the use_path_style parameter, and the default access method is the virtual-hosted style method.
  • access_key: (required)
  • secret_key: (required)
  • use_path_style: (optional) default false . The S3 SDK uses the virtual-hosted style by default. However, some object storage systems may not be enabled or support virtual-hosted style access. At this time, we can add the use_path_style parameter to force the use of path style access method.

Note: URI currently supports three SCHEMA: http://, https:// and s3://.

  1. If you use http:// or https://, you will decide whether to use the ‘path style’ to access s3 based on the ‘use_path_style’ parameter
  2. If you use s3://, you will use the “virtual-hosted style’ to access the s3, ‘use_path_style’ parameter is invalid.

For detailed use cases, you can refer to Best Practice at the bottom.

file format parameter:

  • format: (required) Currently support csv/csv_with_names/csv_with_names_and_types/json/parquet/orc

  • column_separator: (optional) default ,.

  • line_delimiter: (optional) default \n.

    The following 6 parameters are used for loading in json format. For specific usage methods, please refer to: Json Load

  • read_json_by_line: (optional) default "true"

  • strip_outer_array: (optional) default "false"

  • json_root: (optional) default ""

  • json_paths: (optional) default ""

  • num_as_string: (optional) default "false"

  • fuzzy_parse: (optional) default "false"

Example

Read and access csv format files on S3-compatible object storage.

  1. select * from s3("uri" = "http://127.0.0.1:9312/test2/student1.csv",
  2. "ACCESS_KEY"= "minioadmin",
  3. "SECRET_KEY" = "minioadmin",
  4. "format" = "csv",
  5. "use_path_style" = "true") order by c1;

Can be used with desc function

  1. MySQL [(none)]> Desc function s3("uri" = "http://127.0.0.1:9312/test2/student1.csv",
  2. "ACCESS_KEY"= "minioadmin",
  3. "SECRET_KEY" = "minioadmin",
  4. "format" = "csv",
  5. "use_path_style" = "true");

Keywords

  1. s3, table-valued-function, tvf

Best Practice

Since the S3 table-valued-function does not know the table schema in advance, it will read the file first to parse out the table schema.

Usage of different uri schemas Example of http:// 、https://

  1. // Note how to write your bucket of URI and set the 'use_path_style' parameter, as well as http://.
  2. // Because of "use_path_style"="true", s3 will be accessed in 'path style'.
  3. select * from s3(
  4. "URI" = "https://endpoint/bucket/file/student.csv",
  5. "ACCESS_KEY"= "ak",
  6. "SECRET_KEY" = "sk",
  7. "FORMAT" = "csv",
  8. "use_path_style"="true");
  9. // Note how to write your bucket of URI and set the 'use_path_style' parameter, as well as http://.
  10. // Because of "use_path_style"="false", s3 will be accessed in 'virtual-hosted style'.
  11. select * from s3(
  12. "URI" = "https://bucket.endpoint/file/student.csv",
  13. "ACCESS_KEY"= "ak",
  14. "SECRET_KEY" = "sk",
  15. "FORMAT" = "csv",
  16. "use_path_style"="false");
  17. // The OSS on Alibaba Cloud and The COS on Tencent Cloud will use 'virtual-hosted style' to access s3.
  18. // OSS
  19. select * from s3(
  20. "URI" = "http://example-bucket.oss-cn-beijing.aliyuncs.com/your-folder/file.parquet",
  21. "ACCESS_KEY" = "ak",
  22. "SECRET_KEY" = "sk",
  23. "REGION" = "oss-cn-beijing",
  24. "FORMAT" = "parquet",
  25. "use_path_style" = "false");
  26. // COS
  27. select * from s3(
  28. "URI" = "https://example-bucket.cos.ap-hongkong.myqcloud.com/your-folder/file.parquet",
  29. "ACCESS_KEY" = "ak",
  30. "SECRET_KEY" = "sk",
  31. "REGION" = "ap-hongkong",
  32. "FORMAT" = "parquet",
  33. "use_path_style" = "false");

Example of s3://:

  1. // Note how to write your bucket of URI, no need to set 'use_path_style'.
  2. // s3 will be accessed in 'virtual-hosted style'.
  3. select * from s3(
  4. "URI" = "s3://bucket.endpoint/file/student.csv",
  5. "ACCESS_KEY"= "ak",
  6. "SECRET_KEY" = "sk",
  7. "FORMAT" = "csv");

csv foramt csv format: Read the file on S3 and process it as a csv file, read the first line in the file to parse out the table schema. The number of columns in the first line of the file n will be used as the number of columns in the table schema, and the column names of the table schema will be automatically named c1, c2, ..., cn, and the column type is set to String , for example:

The file content of student1.csv:

  1. 1,ftw,12
  2. 2,zs,18
  3. 3,ww,20

use S3 tvf

  1. MySQL [(none)]> select * from s3("uri" = "http://127.0.0.1:9312/test2/student1.csv",
  2. -> "ACCESS_KEY"= "minioadmin",
  3. -> "SECRET_KEY" = "minioadmin",
  4. -> "format" = "csv",
  5. -> "use_path_style" = "true") order by c1;
  6. +------+------+------+
  7. | c1 | c2 | c3 |
  8. +------+------+------+
  9. | 1 | ftw | 12 |
  10. | 2 | zs | 18 |
  11. | 3 | ww | 20 |
  12. +------+------+------+

use desc function S3() to view the table schema

  1. MySQL [(none)]> Desc function s3("uri" = "http://127.0.0.1:9312/test2/student1.csv",
  2. -> "ACCESS_KEY"= "minioadmin",
  3. -> "SECRET_KEY" = "minioadmin",
  4. -> "format" = "csv",
  5. -> "use_path_style" = "true");
  6. +-------+------+------+-------+---------+-------+
  7. | Field | Type | Null | Key | Default | Extra |
  8. +-------+------+------+-------+---------+-------+
  9. | c1 | TEXT | Yes | false | NULL | NONE |
  10. | c2 | TEXT | Yes | false | NULL | NONE |
  11. | c3 | TEXT | Yes | false | NULL | NONE |
  12. +-------+------+------+-------+---------+-------+

csv_with_names foramt csv_with_names format: The first line of the file is used as the number and name of the columns of the table schema, and the column type is set to String, for example:

The file content of student_with_names.csv:

  1. id,name,age
  2. 1,ftw,12
  3. 2,zs,18
  4. 3,ww,20

use S3 tvf

  1. MySQL [(none)]> select * from s3("uri" = "http://127.0.0.1:9312/test2/student_with_names.csv",
  2. -> "ACCESS_KEY"= "minioadmin",
  3. -> "SECRET_KEY" = "minioadmin",
  4. -> "format" = "csv_with_names",
  5. -> "use_path_style" = "true") order by id;
  6. +------+------+------+
  7. | id | name | age |
  8. +------+------+------+
  9. | 1 | ftw | 12 |
  10. | 2 | zs | 18 |
  11. | 3 | ww | 20 |
  12. +------+------+------+
  1. MySQL [(none)]> Desc function s3("uri" = "http://127.0.0.1:9312/test2/student_with_names.csv",
  2. -> "ACCESS_KEY"= "minioadmin",
  3. -> "SECRET_KEY" = "minioadmin",
  4. -> "format" = "csv_with_names",
  5. -> "use_path_style" = "true");
  6. +-------+------+------+-------+---------+-------+
  7. | Field | Type | Null | Key | Default | Extra |
  8. +-------+------+------+-------+---------+-------+
  9. | id | TEXT | Yes | false | NULL | NONE |
  10. | name | TEXT | Yes | false | NULL | NONE |
  11. | age | TEXT | Yes | false | NULL | NONE |
  12. +-------+------+------+-------+---------+-------+

csv_with_names_and_types foramt

csv_with_names_and_types foramt: Currently, it does not support parsing the column type from a csv file. When using this format, S3 tvf will parse the first line of the file as the number and name of the columns of the table schema, and set the column type to String. Meanwhile, the second line of the file is ignored.

The file content of student_with_names_and_types.csv:

  1. id,name,age
  2. INT,STRING,INT
  3. 1,ftw,12
  4. 2,zs,18
  5. 3,ww,20

use S3 tvf

  1. MySQL [(none)]> select * from s3("uri" = "http://127.0.0.1:9312/test2/student_with_names_and_types.csv",
  2. -> "ACCESS_KEY"= "minioadmin",
  3. -> "SECRET_KEY" = "minioadmin",
  4. -> "format" = "csv_with_names_and_types",
  5. -> "use_path_style" = "true") order by id;
  6. +------+------+------+
  7. | id | name | age |
  8. +------+------+------+
  9. | 1 | ftw | 12 |
  10. | 2 | zs | 18 |
  11. | 3 | ww | 20 |
  12. +------+------+------+
  1. MySQL [(none)]> Desc function s3("uri" = "http://127.0.0.1:9312/test2/student_with_names_and_types.csv",
  2. -> "ACCESS_KEY"= "minioadmin",
  3. -> "SECRET_KEY" = "minioadmin",
  4. -> "format" = "csv_with_names_and_types",
  5. -> "use_path_style" = "true");
  6. +-------+------+------+-------+---------+-------+
  7. | Field | Type | Null | Key | Default | Extra |
  8. +-------+------+------+-------+---------+-------+
  9. | id | TEXT | Yes | false | NULL | NONE |
  10. | name | TEXT | Yes | false | NULL | NONE |
  11. | age | TEXT | Yes | false | NULL | NONE |
  12. +-------+------+------+-------+---------+-------+

json foramt

json format: The json format involves many optional parameters, and the meaning of each parameter can be referred to: Json Load. When S3 tvf queries the json format file, it locates a json object according to the json_root and jsonpaths parameters, and uses the key in the object as the column name of the table schema, and sets the column type to String. For example:

The file content of data.json:

  1. [{"id":1, "name":"ftw", "age":18}]
  2. [{"id":2, "name":"xxx", "age":17}]
  3. [{"id":3, "name":"yyy", "age":19}]

use S3 tvf:

  1. MySQL [(none)]> select * from s3(
  2. "URI" = "http://127.0.0.1:9312/test2/data.json",
  3. "ACCESS_KEY"= "minioadmin",
  4. "SECRET_KEY" = "minioadmin",
  5. "Format" = "json",
  6. "strip_outer_array" = "true",
  7. "read_json_by_line" = "true",
  8. "use_path_style"="true");
  9. +------+------+------+
  10. | id | name | age |
  11. +------+------+------+
  12. | 1 | ftw | 18 |
  13. | 2 | xxx | 17 |
  14. | 3 | yyy | 19 |
  15. +------+------+------+
  16. MySQL [(none)]> select * from s3(
  17. "URI" = "http://127.0.0.1:9312/test2/data.json",
  18. "ACCESS_KEY"= "minioadmin",
  19. "SECRET_KEY" = "minioadmin",
  20. "Format" = "json",
  21. "strip_outer_array" = "true",
  22. "jsonpaths" = "[\"$.id\", \"$.age\"]",
  23. "use_path_style"="true");
  24. +------+------+
  25. | id | age |
  26. +------+------+
  27. | 1 | 18 |
  28. | 2 | 17 |
  29. | 3 | 19 |
  30. +------+------+

parquet foramt

parquet format: S3 tvf supports parsing the column names and column types of the table schema from the parquet file. Example:

  1. MySQL [(none)]> select * from s3(
  2. "URI" = "http://127.0.0.1:9312/test2/test.snappy.parquet",
  3. "ACCESS_KEY"= "minioadmin",
  4. "SECRET_KEY" = "minioadmin",
  5. "Format" = "parquet",
  6. "use_path_style"="true") limit 5;
  7. +-----------+------------------------------------------+----------------+----------+-------------------------+--------+-------------+---------------+---------------------+
  8. | p_partkey | p_name | p_mfgr | p_brand | p_type | p_size | p_container | p_retailprice | p_comment |
  9. +-----------+------------------------------------------+----------------+----------+-------------------------+--------+-------------+---------------+---------------------+
  10. | 1 | goldenrod lavender spring chocolate lace | Manufacturer#1 | Brand#13 | PROMO BURNISHED COPPER | 7 | JUMBO PKG | 901 | ly. slyly ironi |
  11. | 2 | blush thistle blue yellow saddle | Manufacturer#1 | Brand#13 | LARGE BRUSHED BRASS | 1 | LG CASE | 902 | lar accounts amo |
  12. | 3 | spring green yellow purple cornsilk | Manufacturer#4 | Brand#42 | STANDARD POLISHED BRASS | 21 | WRAP CASE | 903 | egular deposits hag |
  13. | 4 | cornflower chocolate smoke green pink | Manufacturer#3 | Brand#34 | SMALL PLATED BRASS | 14 | MED DRUM | 904 | p furiously r |
  14. | 5 | forest brown coral puff cream | Manufacturer#3 | Brand#32 | STANDARD POLISHED TIN | 15 | SM PKG | 905 | wake carefully |
  15. +-----------+------------------------------------------+----------------+----------+-------------------------+--------+-------------+---------------+---------------------+
  1. MySQL [(none)]> desc function s3(
  2. "URI" = "http://127.0.0.1:9312/test2/test.snappy.parquet",
  3. "ACCESS_KEY"= "minioadmin",
  4. "SECRET_KEY" = "minioadmin",
  5. "Format" = "parquet",
  6. "use_path_style"="true");
  7. +---------------+--------------+------+-------+---------+-------+
  8. | Field | Type | Null | Key | Default | Extra |
  9. +---------------+--------------+------+-------+---------+-------+
  10. | p_partkey | INT | Yes | false | NULL | NONE |
  11. | p_name | TEXT | Yes | false | NULL | NONE |
  12. | p_mfgr | TEXT | Yes | false | NULL | NONE |
  13. | p_brand | TEXT | Yes | false | NULL | NONE |
  14. | p_type | TEXT | Yes | false | NULL | NONE |
  15. | p_size | INT | Yes | false | NULL | NONE |
  16. | p_container | TEXT | Yes | false | NULL | NONE |
  17. | p_retailprice | DECIMAL(9,0) | Yes | false | NULL | NONE |
  18. | p_comment | TEXT | Yes | false | NULL | NONE |
  19. +---------------+--------------+------+-------+---------+-------+

orc foramt

orc format: Same as parquet format, set format parameter to orc.

  1. MySQL [(none)]> select * from s3(
  2. "URI" = "http://127.0.0.1:9312/test2/test.snappy.orc",
  3. "ACCESS_KEY"= "minioadmin",
  4. "SECRET_KEY" = "minioadmin",
  5. "Format" = "orc",
  6. "use_path_style"="true") limit 5;
  7. +-----------+------------------------------------------+----------------+----------+-------------------------+--------+-------------+---------------+---------------------+
  8. | p_partkey | p_name | p_mfgr | p_brand | p_type | p_size | p_container | p_retailprice | p_comment |
  9. +-----------+------------------------------------------+----------------+----------+-------------------------+--------+-------------+---------------+---------------------+
  10. | 1 | goldenrod lavender spring chocolate lace | Manufacturer#1 | Brand#13 | PROMO BURNISHED COPPER | 7 | JUMBO PKG | 901 | ly. slyly ironi |
  11. | 2 | blush thistle blue yellow saddle | Manufacturer#1 | Brand#13 | LARGE BRUSHED BRASS | 1 | LG CASE | 902 | lar accounts amo |
  12. | 3 | spring green yellow purple cornsilk | Manufacturer#4 | Brand#42 | STANDARD POLISHED BRASS | 21 | WRAP CASE | 903 | egular deposits hag |
  13. | 4 | cornflower chocolate smoke green pink | Manufacturer#3 | Brand#34 | SMALL PLATED BRASS | 14 | MED DRUM | 904 | p furiously r |
  14. | 5 | forest brown coral puff cream | Manufacturer#3 | Brand#32 | STANDARD POLISHED TIN | 15 | SM PKG | 905 | wake carefully |
  15. +-----------+------------------------------------------+----------------+----------+-------------------------+--------+-------------+---------------+---------------------+

uri contains wildcards

uri can use wildcards to read multiple files. Note: If wildcards are used, the format of each file must be consistent (especially csv/csv_with_names/csv_with_names_and_types count as different formats), S3 tvf uses the first file to parse out the table schema. For example:

The following two csv files:

  1. // file1.csv
  2. 1,aaa,18
  3. 2,qqq,20
  4. 3,qwe,19
  5. // file2.csv
  6. 5,cyx,19
  7. 6,ftw,21

You can use wildcards on the uri to query.

  1. MySQL [(none)]> select * from s3(
  2. "URI" = "http://127.0.0.1:9312/test2/file*.csv",
  3. "ACCESS_KEY"= "minioadmin",
  4. "SECRET_KEY" = "minioadmin",
  5. "ForMAT" = "csv",
  6. "use_path_style"="true");
  7. +------+------+------+
  8. | c1 | c2 | c3 |
  9. +------+------+------+
  10. | 1 | aaa | 18 |
  11. | 2 | qqq | 20 |
  12. | 3 | qwe | 19 |
  13. | 5 | cyx | 19 |
  14. | 6 | ftw | 21 |
  15. +------+------+------+

Using S3 tvf with insert into and cast

  1. // Create doris internal table
  2. CREATE TABLE IF NOT EXISTS ${testTable}
  3. (
  4. id int,
  5. name varchar(50),
  6. age int
  7. )
  8. COMMENT "my first table"
  9. DISTRIBUTED BY HASH(id) BUCKETS 32
  10. PROPERTIES("replication_num" = "1");
  11. // Insert data using S3
  12. insert into ${testTable} (id,name,age)
  13. select cast (id as INT) as id, name, cast (age as INT) as age
  14. from s3(
  15. "uri" = "${uri}",
  16. "ACCESS_KEY"= "${ak}",
  17. "SECRET_KEY" = "${sk}",
  18. "format" = "${format}",
  19. "strip_outer_array" = "true",
  20. "read_json_by_line" = "true",
  21. "use_path_style" = "true");