DocsICP FilingConsole
官方文档
首页

UNLOAD

更新时间：
复制 MD 格式
产品详情
我的收藏

The UNLOAD command exports data from a MaxCompute project to external storage, such as Object Storage Service (OSS) or Hologres, for use by other compute engines.

Permissions

  • MaxCompute permissions: The account used for the operation must have read (Select) permission on the table data to be exported. For more information about how to grant permissions, see MaxCompute permissions.

  • External storage permissions: Before you export data from MaxCompute to external storage, you must grant MaxCompute the permissions to access the external storage, which can be OSS or Hologres.

Export data to external storage

The UNLOAD command uses append mode. When you export data to the same destination path multiple times, the system does not overwrite existing files. Instead, it creates new files in the path. To overwrite the data, you must manually clear the destination path before you run the UNLOAD command.

Export data to OSS

Export data in a text format (CSV or TSV)

This method uses a built-in StorageHandler to export data. A .csv or .tsv file name extension is added to the output files by default.

Syntax

UNLOAD FROM {<select_statement>|<table_name> [PARTITION (<pt_spec>)]} 
INTO | OVERWRITE
LOCATION <oss_location>
STORED BY <StorageHandler>
[WITH SERDEPROPERTIES ('<property_name>'='<property_value>',...)];

Core configurations

Configuration item

Description

oss_location

The destination path in OSS. The format is oss://<oss_endpoint>/<bucket>/<object_path>.

StorageHandler

Built-in processor:

  • 'com.aliyun.odps.CsvStorageHandler': exports data in CSV format.

  • 'com.aliyun.odps.TsvStorageHandler': exports data in TSV format.

SERDEPROPERTIES

Used to configure export properties. The most common properties are:

  • 'odps.properties.rolearn': Required. The Alibaba Cloud Resource Name (ARN) of the RAM role that is used to access OSS.

  • 'odps.text.option.gzip.output.enabled'='true': Compresses the exported file in GZIP format.

Examples

Example 1: Export data in CSV format and compress the data into a GZIP file

Export data from the sale_detail table in a MaxCompute project to OSS. The following code shows the data in the sale_detail table:

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    Method 1: Use a SELECT query

    -- Set the amount of data to be read from the MaxCompute table by a single worker to 256 MB. This indirectly controls the number of exported files.
SET odps.stage.mapper.split.size=256;

UNLOAD FROM (SELECT * FROM sale_detail)
INTO
LOCATION 'oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location'
STORED BY 'com.aliyun.odps.CsvStorageHandler'
WITH SERDEPROPERTIES (
    'odps.properties.rolearn'='acs:ram::<uid>:role/AliyunODPSDefaultRole', 
    'odps.text.option.gzip.output.enabled'='true'
);

    Method 2: Directly specify the table name (no query fees are generated)

    -- Set the amount of data to be read from the MaxCompute table by a single worker to 256 MB. This indirectly controls the number of exported files.
SET odps.stage.mapper.split.size=256;

UNLOAD FROM sale_detail 
INTO
LOCATION 'oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location'
STORED BY 'com.aliyun.odps.CsvStorageHandler'
WITH SERDEPROPERTIES (
    'odps.properties.rolearn'='acs:ram::<uid>:role/AliyunODPSDefaultRole', 
    'odps.text.option.gzip.output.enabled'='true'
);

  3. Verify the result: Log on to the OSS console and view the exported data in the destination OSS path. The exported file is in CSV format and is compressed using GZIP.

Example 2: Export partitioned data in TSV format and compress the data

Export the data in the partitions for which sale_date='2013' and region='china' are specified from the sale_detail table in TSV format and compress the data using GZIP.

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    SET odps.stage.mapper.split.size=256;

UNLOAD FROM sale_detail PARTITION (sale_date='2013',region='china')
INTO
LOCATION 'oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location'
STORED BY 'com.aliyun.odps.TsvStorageHandler'
WITH SERDEPROPERTIES (
    'odps.properties.rolearn'='acs:ram::<uid>:role/AliyunODPSDefaultRole', 
    'odps.text.option.gzip.output.enabled'='true'
);

  3. Verify the result: Log on to the OSS Management Console and view the import result in the target OSS path. The exported file is a GZIP-compressed file in the TSV format.

Export data in an open source format (such as Parquet or ORC)

This method supports exporting data to various open-source columnar storage or structured data formats.

Syntax

UNLOAD FROM {<select_statement>|<table_name> [PARTITION (<pt_spec>)]}
INTO | OVERWRITE
LOCATION <oss_location>
[ROW FORMAT SERDE '<serde_class>' 
  [WITH SERDEPROPERTIES ('<property_name>'='<property_value>',...)]
]
STORED AS <file_format>
[PROPERTIES ('<tbproperty_name>'='<tbproperty_value>')];

Core configurations

Configuration item

Description

oss_location

The destination path in OSS. The format is oss://<oss_endpoint>/<bucket>/<object_path>.

serde_class

Specifies the serialization and deserialization library. For example, use org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe for Parquet.

SERDEPROPERTIES

'odps.properties.rolearn': Required. The ARN of the RAM role that is used to access OSS.

file_format

Required. Specifies the file format, such as PARQUET, ORC, TEXTFILE, or JSON.

PROPERTIES

- 'mcfed.parquet.compression'='SNAPPY': Specifies the compression format for Parquet files. SNAPPY and LZO are supported.
- 'odps.text.option.gzip.output.enabled'='true': Specifies GZIP compression for TextFile.

Supported formats and compression methods

File format

Supported compression formats

Description

PARQUET

SNAPPY, LZO

A columnar storage format that is suitable for analytical queries.

ORC

SNAPPY, LZO

A columnar storage format that is suitable for the Hadoop ecosystem.

TEXTFILE

GZIP

A text format that supports custom separators.

RCFILE

-

A row-column hybrid storage format.

SEQUENCEFILE

-

A Hadoop sequence file format.

AVRO

-

A data serialization format.

JSON

-

A JSON format.

Examples

Example 1: Export data in PARQUET format and compress the data using SNAPPY

Export data from the sale_detail table in a MaxCompute project to OSS. The following code shows the data in the sale_detail table:

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    -- Set the amount of data to be read from the MaxCompute table by a single worker to 256 MB. This indirectly controls the number of exported files.
SET odps.stage.mapper.split.size=256;

UNLOAD FROM (SELECT * FROM sale_detail) 
INTO  
LOCATION 'oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location' 
ROW FORMAT SERDE 'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe' 
WITH SERDEPROPERTIES (
    'odps.properties.rolearn'='acs:ram::<uid>:role/AliyunODPSDefaultRole'
) 
STORED AS PARQUET 
PROPERTIES('mcfed.parquet.compression'='SNAPPY');

  3. Verification Result: Log on to the OSS Management Console and view the import result in the target OSS path. The exported files are in PARQUET format and compressed using SNAPPY.

Example 2: Export partitioned data in PARQUET format

Export the data in the partitions for which sale_date='2013' and region='china' are specified from the MaxCompute sale_detail table in PARQUET format and compress the data using SNAPPY. The following code shows the data in the sale_detail table:

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    -- Set the amount of data to be read from the MaxCompute table by a single worker to 256 MB. This indirectly controls the number of exported files.
SET odps.stage.mapper.split.size=256;

UNLOAD FROM sale_detail PARTITION (sale_date='2013',region='china') 
INTO 
LOCATION 'oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location' 
ROW FORMAT SERDE 'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe' 
WITH SERDEPROPERTIES ('odps.properties.rolearn'='acs:ram::<uid>:role/AliyunODPSDefaultRole') 
STORED AS PARQUET 
PROPERTIES('mcfed.parquet.compression'='SNAPPY');

  3. Verify the result: Log on to the OSS console and view the exported data in the destination OSS path. The exported file is in PARQUET format and is compressed using SNAPPY.

Example 3: Export data in TEXTFILE format and specify a separator

Export data from the sale_detail table to a TXT file and use a comma (,) as the separator.

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    -- Set the amount of data to be read from the MaxCompute table by a single worker to 256 MB. This indirectly controls the number of exported files.
SET odps.sql.allow.fullscan=true; 

UNLOAD FROM (SELECT * FROM sale_detail)
INTO
LOCATION 'oss://oss-cn-beijing-internal.aliyuncs.com/mc-unload/data_location/'
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
WITH SERDEPROPERTIES ('field.delim'=',')
STORED AS TEXTFILE
PROPERTIES ('odps.external.data.enable.extension'='true');

  3. Verify the result: Log on to the OSS console and view the exported data in the destination OSS path. The exported file is in TXT format and uses commas as separators.

Customize file name prefixes, suffixes, and extensions

When you use the UNLOAD command to export data to OSS, you can use properties in PROPERTIES or SERDEPROPERTIES to customize the prefix, suffix, and extension of output files.

Syntax

UNLOAD FROM {<select_statement>|<table_name> [PARTITION (<pt_spec>)]}
INTO | OVERWRITE
LOCATION <oss_location>
[ROW FORMAT SERDE '<serde_class>' 
  [WITH SERDEPROPERTIES ('<property_name>'='<property_value>',...)]
]
STORED AS <file_format>
[PROPERTIES ('<tbproperty_name>'='<tbproperty_value>',...)];

Core configurations

property_name/tbproperty_name

Description

Example value

odps.external.data.output.prefix

The prefix of the file name. It can contain letters, digits, and underscores (_). The prefix can be 1 to 10 characters in length.

'mc_'

odps.external.data.output.suffix

The suffix of the file name. It can contain letters, digits, and underscores (_).

'_hangzhou'

odps.external.data.enable.extension

Specifies whether to show default extensions such as .tx and .parquet.

'true'

odps.external.data.output.explicit.extension

A custom file name extension. This property has a higher priority than the default extension.

'jsonl'

Suffix reference

The following table describes the extensions that are automatically generated for different foreign tables when you set the odps.external.data.enable.extension=true parameter.

File format

SerDe

Suffix

SEQUENCEFILE

org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

.sequencefile

TEXTFILE

org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

.txt

RCFILE

org.apache.hadoop.hive.serde2.columnar.LazyBinaryColumnarSerDe

.rcfile

ORC

org.apache.hadoop.hive.ql.io.orc.OrcSerde

.orc

PARQUET

org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe

.parquet

AVRO

org.apache.hadoop.hive.serde2.avro.AvroSerDe

.avro

JSON

org.apache.hive.hcatalog.data.JsonSerDe

.json

CSV

org.apache.hadoop.hive.serde2.OpenCSVSerde

.csv

Note

When you export data that is compressed in SNAPPY or LZO format, the .snappy or .lzo extension is not displayed in the names of the exported files.

Examples

Example 1: Export a TEXTFILE format file and add a prefix and suffix

Export data from the sale_detail table in a MaxCompute project to OSS as a TXT file. The file is named in the mc_<system-generated>_beijing.txt format. The following code shows the data in the sale_detail table:

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    UNLOAD FROM (SELECT * FROM sale_detail) 
INTO 
LOCATION 'oss://oss-cn-beijing-internal.aliyuncs.com/***/textfile' 
row format serde 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
STORED AS textfile
PROPERTIES (
    'odps.external.data.output.prefix'='mc_', 
    'odps.external.data.output.suffix'='_beijing',
    'odps.external.data.enable.extension'='true');

  3. Verify the result: Log on to the OSS console and view the exported data in the destination OSS path. The name of the exported file is mc_<system-generated>_beijing.txt.

Example 2: Export a JSON format file and customize the file name extension

Export data from the sale_detail table in a MaxCompute project to OSS as a JSON file. The file is named in the mc_<system-generated file name>_beijing.json format. The following code shows the data in the sale_detail table:

-- Partition fields: sale_date, region
+------------+-------------+-------------+------------+------------+
| shop_name  | customer_id | total_price | sale_date  | region     |
+------------+-------------+-------------+------------+------------+
| s1         | c1          | 100.1       | 2013       | china      |
| s2         | c2          | 100.2       | 2013       | china      |
| s3         | c3          | 100.3       | 2013       | china      |
| null       | c5          | NULL        | 2014       | shanghai   |
| s6         | c6          | 100.4       | 2014       | shanghai   |
| s7         | c7          | 100.5       | 2014       | shanghai   |
+------------+-------------+-------------+------------+------------+

Procedure:

  1. On the OSS side: Log on to the OSS console, create the mc-unload/data_location/ folder in an OSS bucket that is in the oss-cn-hangzhou region, and then construct an OSS path. For more information about how to create an OSS bucket, see Create a bucket in the console. Construct the OSS path based on the bucket, region, and endpoint information. For example:

    oss://oss-cn-hangzhou-internal.aliyuncs.com/mc-unload/data_location

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    UNLOAD FROM (SELECT * FROM sale_detail) 
INTO 
LOCATION 'oss://oss-cn-beijing-internal.aliyuncs.com/***/json' 
ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
WITH SERDEPROPERTIES (
    'odps.external.data.output.prefix'='mc_', 
    'odps.external.data.output.suffix'='_beijing',
    'odps.external.data.output.explicit.extension'='json')
STORED AS JSON;

  3. Verify the result: Log on to the OSS console and view the exported data in the destination OSS path. The name of the exported file is mc_<system-generated file name>_beijing.json.

Export data to Hologres

Limits

  • Dual-signature authorization: Dual-signature authorization is not supported for exporting data to Hologres.

  • Partitioned tables: You cannot export data to Hologres partitioned tables.

  • Type mapping: The data types of fields in the destination Hologres table must correspond to the data types of fields in the MaxCompute table. For more information, see Data type mapping between MaxCompute and Hologres.

Syntax

UNLOAD FROM {<select_statement>|<table_name> [PARTITION (<pt_spec>)]} 
INTO 
LOCATION <hologres_location>
STORED BY <StorageHandler>
[WITH SERDEPROPERTIES ('<property_name>'='<property_value>',...)];

Core configurations

Configuration item

Description

hologres_location

The Java Database Connectivity (JDBC) connection string of the destination Hologres table. The format is 'jdbc:postgresql://<endpoint>:<port>/<database>?ApplicationName=MaxCompute&[currentSchema=<schema>&][useSSL={true|false}&]table=<holo_table_name>/'.

StorageHandler

Specifies a built-in handler. The value is fixed at com.aliyun.odps.jdbc.JdbcStorageHandler, which uses the JdbcStorageHandler connection method.

SERDEPROPERTIES

This parameter must contain the following three properties:

  • 'odps.properties.rolearn'='<ram_arn>': The ARN of the RAM role that is used to access Hologres.

  • 'mcfed.mapreduce.jdbc.driver.class'='org.postgresql.Driver': The JDBC driver of Hologres.

  • 'odps.federation.jdbc.target.db.type'='holo': The type of the destination database.

Example

Export data from the MaxCompute table data_test to the Hologres table mc_2_holo. The following code shows the data in the data_test table:

+------------+------+
| id         | name |
+------------+------+
| 3          | rgege |
| 4          | Gegegegr |
+------------+------+

Procedure:

  1. On the Hologres side: Create the destination table mc_2_holo in Hologres.

    CREATE TABLE mc_2_holo (id INT, name TEXT);

  2. On the MaxCompute side: Log on to a MaxCompute client and run the UNLOAD command.

    UNLOAD FROM (SELECT * FROM data_test) 
INTO 
LOCATION 'jdbc:postgresql://hgprecn-cn-5y**********-cn-hangzhou-internal.hologres.aliyuncs.com:80/test?ApplicationName=MaxCompute&currentSchema=public&useSSL=false&table=mc_2_holo/'  
STORED BY 'com.aliyun.odps.jdbc.JdbcStorageHandler'
WITH SERDEPROPERTIES ( 
  'odps.properties.rolearn'='acs:ram::13**************:role/aliyunodpsholorole',
  'mcfed.mapreduce.jdbc.driver.class'='org.postgresql.Driver', 
  'odps.federation.jdbc.target.db.type'='holo'
);

  3. Verify the result: View the exported data in Hologres.

    SELECT * FROM mc_2_holo;

    The following result is returned:

    id	name
4	Gegegegr
3	rgege

Billing

  • Command billing

    • The UNLOAD command itself is not billed.

    • Billing for query clauses: The query clause in an UNLOAD command scans data and uses computing resources to generate the exported data. Therefore, the query clause is billed as a standard SQL job.

  • Cost estimation

    Storing structured data in OSS and storing data in MaxCompute each have their own advantages. To reduce costs when you export data from MaxCompute, perform a comprehensive cost estimation that includes direct storage costs, data volume changes, and subsequent query costs.

    • Direct storage costs

      • MaxCompute provides the Standard, Infrequent Access (IA), and Archive storage classes. All storage classes support zone-redundant storage. For more information about storage billing, see Storage fees.

      • OSS also provides Standard, IA, and Archive Storage classes, all of which support zone-redundant storage. The price of each storage class is the same as that in MaxCompute. OSS also provides Cold Archive and Deep Cold Archive storage classes that are less expensive but do not support zone-redundant storage. For more information, see OSS storage fees.

    • Data volume changes

      Data imported into MaxCompute has a compression ratio of approximately 5:1. Billing is based on the amount of data after compression.

      For data stored in OSS, you must specify and maintain the data format and compression method. For more information about the data formats and compression methods that are supported for exporting data from MaxCompute, see OSS foreign tables. If data is not compressed, the data volume expands after the data is exported to OSS.

    • Subsequent query costs

      Data stored in MaxCompute can be queried at any time. However, if you use storage classes such as Infrequent Access or Archive, you must pay additional data access fees for reading data. If data is stored in non-standard OSS storage classes such as Infrequent Access, Archive, or Cold Archive, data retrieval fees are also incurred. If the data is deleted or retrieved before the minimum storage period required for the storage class is met, you may need to pay early deletion fees. For more information, see the pricing calculator.

References

To import data in CSV or other open-source formats from external storage into MaxCompute, see LOAD.

Previous：UNINSTALL PACKAGENext：UPDATE