COPY INTO (Transact-SQL)

Applies to: Azure Synapse Analytics

Tip

Microsoft Fabric Data Warehouse is an enterprise scale relational warehouse on a data lake foundation, with a future-ready architecture, built-in AI, and new features. If you're new to data warehousing, start with Fabric Data Warehouse. Existing dedicated SQL pool workloads can upgrade to Fabric to access new capabilities across data science, real-time analytics, and reporting.

• Start a Fabric free trial.
• Migration Assistant for Fabric Data Warehouse.

This article explains how to use the COPY statement in Azure Synapse Analytics for loading data from external storage accounts. The COPY statement provides the most flexibility for high-throughput data ingestion into Azure Synapse Analytics.

Note

For Warehouse in Microsoft Fabric, see COPY INTO.

Use COPY for the following capabilities:

• Use lower privileged users to load data without needing strict CONTROL permissions on the data warehouse.
• Execute a single T-SQL statement without having to create any other database objects.
• Properly parse and load CSV files where delimiters (string, field, row) are escaped within string delimited columns.
• Specify a finer permission model without exposing storage account keys by using Shared Access Signatures (SAS).
• Use a different storage account for the ERRORFILE location (REJECTED_ROW_LOCATION).
• Customize default values for each target column and specify source data fields to load into specific target columns.
• Specify a custom row terminator, field terminator, and field quote for CSV files.
• Use SQL Server Date formats for CSV files.
• Specify wildcards and multiple files in the storage location path.
• Automatic schema discovery simplifies the process of defining and mapping source data into target tables.
• The automatic table creation process automatically creates the tables and works alongside automatic schema discovery.
• Directly load complex data types from Parquet files such as Maps and Lists into string columns, without using other tools to preprocess the data.

Note

To load complex data types from Parquet files, turn on automatic table creation by using AUTO_CREATE_TABLE.

For comprehensive examples and quickstarts using the COPY statement, see:

• Quickstart: Bulk load data using the COPY statement
• Quickstart: Examples using the COPY statement and its supported authentication methods
• Quickstart: Creating the COPY statement using the rich Synapse Studio UI

Note

Microsoft Entra ID was previously known as Azure Active Directory (Azure AD).

Syntax

COPY INTO [ schema. ] table_name
[ (Column_list) ]
FROM '<external_location>' [ , ...n ]
WITH
 (
 [ FILE_TYPE = { 'CSV' | 'PARQUET' | 'ORC' } ]
 [ , FILE_FORMAT = EXTERNAL FILE FORMAT OBJECT ]
 [ , CREDENTIAL = (AZURE CREDENTIAL) ]
 [ , ERRORFILE = ' [ http(s)://storageaccount/container ] /errorfile_directory [ / ] ] '
 [ , ERRORFILE_CREDENTIAL = (AZURE CREDENTIAL) ]
 [ , MAXERRORS = max_errors ]
 [ , COMPRESSION = { 'Gzip' | 'DefaultCodec' | 'Snappy' } ]
 [ , FIELDQUOTE = 'string_delimiter' ]
 [ , FIELDTERMINATOR =  'field_terminator' ]
 [ , ROWTERMINATOR = 'row_terminator' ]
 [ , FIRSTROW = first_row ]
 [ , DATEFORMAT = 'date_format' ]
 [ , ENCODING = { 'UTF8' | 'UTF16' } ]
 [ , IDENTITY_INSERT = { 'ON' | 'OFF' } ]
 [ , AUTO_CREATE_TABLE = { 'ON' | 'OFF' } ]
)

Arguments

schema_name

Optional if the default schema for the user performing the operation is the schema of the specified table. If you don't specify schema, and the default schema of the user performing the COPY operation is different from the schema of the specified table, the COPY operation is canceled and an error message is returned.

table_name

The name of the table to COPY data into. The target table can be a temporary or permanent table and must already exist in the database. For automatic schema detection mode, don't provide a column list.

(column_list)

An optional list of one or more columns used to map source data fields to target table columns for loading data.

Don't specify a column_list when AUTO_CREATE_TABLE = 'ON'.

column_list must be enclosed in parentheses and delimited by commas. The column list is of the following format:

[(Column_name [default Default_value] [Field_number] [,...n])]

• Column_name - the name of the column in the target table.
• Default_value - the default value that replaces any NULL value in the input file. Default value applies to all file formats. COPY attempts to load NULL from the input file when a column is omitted from the column list or when there's an empty input file field. Default value precedes the keyword 'default'
• Field_number - the input file field number that is mapped to the target column.
• The field indexing starts at 1.

When you don't specify a column list, COPY maps columns based on the source and target order: Input field 1 goes to target column 1, field 2 goes to column 2, and so on.

External locations

The location where the files containing the data are staged. Currently, Azure Data Lake Storage (ADLS) Gen2 and Azure Blob Storage are supported:

• External location for Blob Storage: https://<account\>.blob.core.windows.net/<container\>/<path\>
• External location for ADLS Gen2: https://<account\>.dfs.core.windows.net/<container\>/<path\>

Note

The .blob endpoint is available for ADLS Gen2 as well and currently yields the best performance. Use the .blob endpoint when .dfs isn't required for your authentication method.

• Account - The storage account name

• Container - The blob container name

• Path - the folder or file path for the data. The location starts from the container. If you specify a folder, COPY retrieves all files from the folder and all its subfolders. COPY ignores hidden folders and doesn't return files that begin with an underline (_) or a period (.) unless explicitly specified in the path. This behavior is the same even when specifying a path with a wildcard.

You can include wildcards in the path where:

• Wildcard path name matching is case-sensitive
• You can escape a wildcard by using the backslash character (\)
• Wildcard expansion is applied recursively. For instance, all CSV files under Customer1 (including subdirectories of Customer1) are loaded in the following example: Account/Container/Customer1/*.csv

Note

For best performance, avoid specifying wildcards that expand over a larger number of files. If possible, list multiple file locations instead of specifying wildcards.

You can specify multiple file locations only from the same storage account and container through a comma-separated list such as:

• https://<account>.blob.core.windows.net/<container\>/<path\>, https://<account\>.blob.core.windows.net/<container\>/<path\>

FILE_TYPE = { 'CSV' | 'PARQUET' | 'ORC' }

FILE_TYPE specifies the format of the external data.

• CSV: Specifies a comma-separated values file compliant with the RFC 4180 standard.
• PARQUET: Specifies a Parquet format.
• ORC: Specifies an Optimized Row Columnar (ORC) format.

Note

The 'Delimited Text' file type in PolyBase is replaced by the 'CSV' file format. You can configure the default comma delimiter through the FIELDTERMINATOR parameter.

FILE_FORMAT = external_file_format_name

FILE_FORMAT applies to Parquet and ORC files only. It specifies the name of the external file format object that stores the file type and compression method for the external data. To create an external file format, use CREATE EXTERNAL FILE FORMAT.

CREDENTIAL (IDENTITY = '', SECRET = '')

CREDENTIAL specifies the authentication mechanism to access the external storage account. Authentication methods are:

	CSV	Parquet	ORC
Azure Blob Storage	SAS/MSI/SERVICE PRINCIPAL/KEY/Entra	SAS/KEY	SAS/KEY
Azure Data Lake Gen2	SAS/MSI/SERVICE PRINCIPAL/KEY/Entra	SAS (blob 1 )/MSI (dfs 2 )/SERVICE PRINCIPAL/KEY/Entra	SAS (blob 1 )/MSI (dfs 2 )/SERVICE PRINCIPAL/KEY/Entra

¹ The blob endpoint (.blob.core.windows.net) in your external location path is required for this authentication method.

² The dfs endpoint (.dfs.core.windows.net) in your external location path is required for this authentication method.

Note

• When authenticating by using Microsoft Entra ID or to a public storage account, you don't need to specify CREDENTIAL.
• If your storage account is associated with a VNet, you must authenticate by using a managed identity.

• Authenticating with Shared Access Signatures (SAS)

  • IDENTITY: A constant with a value of Shared Access Signature
  • SECRET: The shared access signature provides delegated access to resources in your storage account.

• Minimum permissions required: READ and LIST

• Authenticating with Service Principals

  • IDENTITY: <ClientID>@<OAuth_2.0_Token_EndPoint>
  • SECRET: Microsoft Entra application service principal key

• Minimum RBAC roles required: Storage blob data contributor, Storage blob data contributor, Storage blob data owner, or Storage blob data reader

• Authenticating with Storage account key

  • IDENTITY: A constant with a value of Storage Account Key
  • SECRET: Storage account key

• Authenticating with Managed Identity (VNet Service Endpoints)

  • IDENTITY: A constant with a value of Managed Identity

• Minimum RBAC roles required: Storage blob data contributor or Storage blob data owner for the Microsoft Entra registered logical server in Azure. When using a dedicated SQL pool (formerly SQL DW) that isn't associated with a Synapse Workspace, this RBAC role isn't required, but the managed identity requires Access Control List (ACL) permissions on the target objects to enable read access to the source files.

• Authenticating with a Microsoft Entra user

  • CREDENTIAL isn't required

• Minimum RBAC roles required: Storage blob data contributor or Storage blob data owner for the Microsoft Entra user

ERRORFILE = Directory Location

ERRORFILE only applies to CSV. It specifies the directory within the COPY statement where the rejected rows and the corresponding error file are written. You can specify the full path from the storage account or the path relative to the container. If the specified path doesn't exist, the warehouse creates one. A child directory is created with the name _rejectedrows. The _ character ensures that the directory is escaped for other data processing unless explicitly named in the location parameter.

Note

When you pass a relative path to ERRORFILE, make it relative to the container path you specify in external_location.

Within this directory, the warehouse creates a folder based on the time of load submission in the format YearMonthDay -HourMinuteSecond (for example, 20180330-173205). In this folder, the process writes two types of files: the reason (error) file and the data (row) file. Each file prepends the queryID, distributionID, and a file GUID. Because the data and the reason are in separate files, corresponding files have a matching prefix.

If ERRORFILE has the full path of the storage account defined, COPY uses ERRORFILE_CREDENTIAL to connect to that storage. Otherwise, it uses the value you specify for CREDENTIAL. When you use the same credential for the source data and ERRORFILE, restrictions that apply to ERRORFILE_CREDENTIAL also apply.

ERRORFILE_CREDENTIAL = (IDENTITY = '', SECRET = '')

ERRORFILE_CREDENTIAL only applies to CSV files. Supported data source and authentication methods are:

• Azure Blob Storage: SAS, service principal, or Microsoft Entra

• Azure Data Lake Gen2: SAS, MSI, service principal, or Microsoft Entra

• Authenticating with Shared Access Signatures (SAS)

  • IDENTITY: A constant with a value of Shared Access Signature
  • SECRET: The shared access signature provides delegated access to resources in your storage account.

• Minimum permissions required: READ, LIST, WRITE, CREATE, DELETE

• Authenticating with Service Principals

  • IDENTITY: <ClientID>@<OAuth_2.0_Token_EndPoint>
  • SECRET: Microsoft Entra application service principal key

• Minimum RBAC roles required: Storage blob data contributor or Storage blob data owner

Note

Use the OAuth 2.0 token endpoint V1

• Authenticating with Managed Identity (VNet Service Endpoints)

  • IDENTITY: A constant with a value of Managed Identity

• Minimum RBAC roles required: Storage blob data contributor or Storage blob data owner for the Microsoft Entra registered SQL Database server

• Authenticating with a Microsoft Entra user

  • CREDENTIAL isn't required

• Minimum RBAC roles required: Storage blob data contributor or Storage blob data owner for the Microsoft Entra user

Using a storage account key with ERRORFILE_CREDENTIAL isn't supported.

Note

If you use the same storage account for your error file and specify the ERRORFILE path relative to the root of the container, you don't need to specify the ERROR_CREDENTIAL.

MAXERRORS = max_errors

MAXERRORS specifies the maximum number of reject rows allowed in the load before the COPY operation fails. Each row that the COPY operation can't import is ignored and counted as one error. If you don't specify a value for maximum number of errors, the default is 0.

MAXERRORS can't be used with AUTO_CREATE_TABLE.

When FILE_TYPE is PARQUET, exceptions that are caused by data type conversion errors (for example, Parquet binary to SQL integer) still cause COPY INTO to fail, ignoring MAXERRORS.

COMPRESSION = { 'DefaultCodec ' | 'Snappy' | 'GZIP' | 'NONE'}

COMPRESSION is optional and specifies the data compression method for the external data.

• CSV supports GZIP.
• Parquet supports GZIP and Snappy.
• ORC supports DefaultCodec and Snappy.
• Zlib is the default compression for ORC.

The COPY command autodetects the compression type based on the file extension when you don't specify this parameter:

• .gz - GZIP
• .snappy - Snappy
• .deflate - DefaultCodec (Parquet and ORC only)

The COPY command requires that gzip files don't contain any trailing garbage to operate normally. The gzip format strictly requires that files be composed of valid members without any additional information before, between, or after them. Any deviation from this format, such as the presence of trailing non-gzip data, results in the failure of the COPY command. To ensure COPY runs successfully, make sure to verify there's no trailing garbage at the end of gzip files.

FIELDQUOTE = 'field_quote'

FIELDQUOTE applies to CSV and specifies a single character that's used as the quote character (string delimiter) in the CSV file. If you don't specify this value, the quote character (") is used as the quote character as defined in the RFC 4180 standard. Hexadecimal notation is also supported for FIELDQUOTE. Extended ASCII and multibyte characters aren't supported with UTF-8 for FIELDQUOTE.

Note

FIELDQUOTE characters are escaped in string columns where there's a presence of a double FIELDQUOTE (delimiter).

FIELDTERMINATOR = 'field_terminator'

FIELDTERMINATOR only applies to CSV. Specifies the field terminator that's used in the CSV file. You can specify the field terminator by using hexadecimal notation. The field terminator can be multicharacter. The default field terminator is a (,). Extended ASCII and multibyte characters aren't supported with UTF-8 for FIELDTERMINATOR.

ROWTERMINATOR = 'row_terminator'

ROWTERMINATOR only applies to CSV. Specifies the row terminator that's used in the CSV file. You can specify the row terminator by using hexadecimal notation. The row terminator can be multicharacter. By default, the row terminator is \r\n.

The COPY command prefixes the \r character when specifying \n (newline) resulting in \r\n. To specify only the \n character, use hexadecimal notation (0x0A). When specifying multicharacter row terminators in hexadecimal, don't specify 0x between each character.

Extended ASCII and multibyte characters aren't supported with UTF-8 for ROWTERMINATOR.

FIRSTROW = First_row_int

FIRSTROW applies to CSV and specifies the row number that's read first in all files for the COPY command. Values start from 1, which is the default value. If you set the value to 2, the first row in every file (header row) is skipped when the data is loaded. Rows are skipped based on the existence of row terminators.

DATEFORMAT = { 'mdy' | 'dmy' | 'ymd' | 'ydm' | 'myd' | 'dym' }

DATEFORMAT only applies to CSV and specifies the date format of the date mapping to SQL Server date formats. For an overview of all Transact-SQL date and time data types and functions, see Date and Time Data Types and Functions (Transact-SQL). DATEFORMAT within the COPY command takes precedence over DATEFORMAT configured at the session level.

ENCODING = 'UTF8' | 'UTF16'

ENCODING only applies to CSV. Default is UTF8. Specifies the data encoding standard for the files loaded by the COPY command.

IDENTITY_INSERT = 'ON' | 'OFF'

IDENTITY_INSERT specifies whether the identity value or values in the imported data file are to be used for the identity column. If IDENTITY_INSERT is OFF (default), the identity values for this column are verified, but not imported. Note the following behavior with the COPY command:

• If IDENTITY_INSERT is OFF, and table has an identity column
  • You must specify a column list that doesn't map an input field to the identity column.
• If IDENTITY_INSERT is ON, and table has an identity column
  • If you pass a column list, it must map an input field to the identity column.
• Default value isn't supported for the IDENTITY column in the column list.
• You can set IDENTITY_INSERT for only one table at a time.

Azure Synapse Analytics automatically assigns unique values based on the seed and increment values specified during table creation.

AUTO_CREATE_TABLE = { 'ON' | 'OFF' }

AUTO_CREATE_TABLE specifies if the table can be automatically created by working alongside automatic schema discovery. AUTO_CREATE_TABLE is available only for Parquet files in Azure Synapse Analytics.

• ON: Enables automatic table creation. The COPY INTO process automatically creates a new table by discovering the structure of the file to be loaded. You can also use it with preexisting tables to take advantage of automatic schema discovery of Parquet files.
• OFF: Automatic table creation isn't enabled. Default.

Note

The automatic table creation works alongside automatic schema discovery. The automatic table creation isn't enabled by default.

Permissions

The user running the COPY command must have the following permissions:

• ADMINISTER DATABASE BULK OPERATIONS
• INSERT

Requires INSERT and ADMINISTER BULK OPERATIONS permissions. In Azure Synapse Analytics, INSERT, and ADMINISTER DATABASE BULK OPERATIONS permissions are required.

Additionally, if the user executing the COPY command also intends to generate a new table and load data into it, they require CREATE TABLE and ALTER ON SCHEMA permissions.

For example, to allow mike@contoso.com to use COPY to create a new table in the HR schema, and insert the data from a Parquet file, use the following Transact-SQL sample:

GRANT ADMINISTER DATABASE BULK OPERATIONS to [mike@contoso.com];
GRANT INSERT to [mike@contoso.com];

GRANT CREATE TABLE to [mike@contoso.com];
GRANT ALTER on SCHEMA::HR to [mike@contoso.com];

Remarks

The COPY statement accepts only UTF-8 and UTF-16 valid characters for row data and command parameters. The COPY statement might incorrectly interpret source files or parameters (such as ROWTERMINATOR or FIELDTERMINATOR) that use invalid characters and cause unexpected results such as data corruption or other failures. Make sure your source files and parameters are UTF-8 or UTF-16 compliant before you invoke the COPY statement.

The MAXDOP query hint isn't supported with COPY INTO.

To ensure reliable execution, don't change the source files and folders during the COPY INTO operation.

• Modifying, deleting, or replacing any referenced files or folders while the command is running can cause the operation to fail or result in inconsistent data ingestion.
• Before executing COPY INTO, verify that all source data is stable and won't be altered during the process.

If the source data has greater precision than the destination column definition, the value is truncated, not rounded, for numeric, date, and time types.

Examples

A. Load from a public storage account

The following example shows the simplest form of the COPY command, which loads data from a public storage account. For this example, the COPY statement's defaults match the format of the line item CSV file.

COPY INTO dbo.[lineitem]
FROM 'https://unsecureaccount.blob.core.windows.net/customerdatasets/folder1/lineitem.csv'
WITH (FIELDTERMINATOR = '|')

The default values of the COPY command are:

• DATEFORMAT = Session DATEFORMAT

• MAXERRORS = 0

• COMPRESSION default is uncompressed

• FIELDQUOTE = '"'

• FIELDTERMINATOR = ','

• ROWTERMINATOR = '\n'

Important

COPY treats \n as \r\n internally. For more information, see the ROWTERMINATOR section.

• FIRSTROW = 1

• ENCODING = 'UTF8'

• FILE_TYPE = 'CSV'

• IDENTITY_INSERT = 'OFF'

B. Load authenticating via Share Access Signature (SAS)

The following example loads files that use the line feed as a row terminator, such as a UNIX output. This example also uses a SAS key to authenticate to Azure Blob Storage.

COPY INTO test_1
FROM 'https://myaccount.blob.core.windows.net/myblobcontainer/folder1/'
WITH (
    FILE_TYPE = 'CSV',
    CREDENTIAL=(IDENTITY= 'Shared Access Signature', SECRET='<Your_SAS_Token>'),
    --CREDENTIAL should look something like this:
    --CREDENTIAL=(IDENTITY= 'Shared Access Signature', SECRET='?sv=2018-03-28&ss=bfqt&srt=sco&sp=rl&st=2016-10-17T20%3A14%3A55Z&se=2021-10-18T20%3A19%3A00Z&sig=IEoOdmeYnE9%2FKiJDSHFSYsz4AkNa%2F%2BTx61FuQ%2FfKHefqoBE%3D'),
    FIELDQUOTE = '"',
    FIELDTERMINATOR=';',
    ROWTERMINATOR='0X0A',
    ENCODING = 'UTF8',
    DATEFORMAT = 'ymd',
    MAXERRORS = 10,
    ERRORFILE = '/errorsfolder',--path starting from the storage container
    IDENTITY_INSERT = 'ON'
)

C. Load with a column list with default values authenticating via Storage Account Key

This example loads files specifying a column list with default values.

--Note when specifying the column list, input field numbers start from 1
COPY INTO test_1 (Col_one default 'myStringDefault' 1, Col_two default 1 3)
FROM 'https://myaccount.blob.core.windows.net/myblobcontainer/folder1/'
WITH (
    FILE_TYPE = 'CSV',
    CREDENTIAL=(IDENTITY= 'Storage Account Key', SECRET='<Your_Account_Key>'),
    FIELDQUOTE = '"',
    FIELDTERMINATOR=',',
    ROWTERMINATOR='0x0A',
    ENCODING = 'UTF8',
    FIRSTROW = 2
)

D. Load Parquet or ORC using existing file format object

This example uses a wildcard to load all Parquet files under a folder.

COPY INTO test_parquet
FROM 'https://myaccount.blob.core.windows.net/myblobcontainer/folder1/*.parquet'
WITH (
    FILE_FORMAT = myFileFormat,
    CREDENTIAL=(IDENTITY= 'Shared Access Signature', SECRET='<Your_SAS_Token>')
)

E. Load specifying wild cards and multiple files

COPY INTO t1
FROM
'https://myaccount.blob.core.windows.net/myblobcontainer/folder0/*.txt',
    'https://myaccount.blob.core.windows.net/myblobcontainer/folder1'
WITH (
    FILE_TYPE = 'CSV',
    CREDENTIAL=(IDENTITY= '<client_id>@<OAuth_2.0_Token_EndPoint>',SECRET='<key>'),
    FIELDTERMINATOR = '|'
)

F. Load using MSI credentials

COPY INTO dbo.myCOPYDemoTable
FROM 'https://myaccount.blob.core.windows.net/myblobcontainer/folder0/*.txt'
WITH (
    FILE_TYPE = 'CSV',
    CREDENTIAL = (IDENTITY = 'Managed Identity'),
    FIELDQUOTE = '"',
    FIELDTERMINATOR=','
)

G. Load using automatic schema detection

COPY INTO [myCOPYDemoTable]
FROM 'https://myaccount.blob.core.windows.net/customerdatasets/folder1/lineitem.parquet'
WITH (
    FILE_TYPE = 'Parquet',
    CREDENTIAL = ( IDENTITY = 'Shared Access Signature',  SECRET='<key>'),
    AUTO_CREATE_TABLE = 'ON'
)

FAQ

How does the performance of the COPY command compare to PolyBase?

The performance of the COPY command can be better depending on your workload.

• The warehouse can't automatically split compressed files. For the best loading performance, consider splitting your input into multiple files when loading compressed CSVs.

• The warehouse can automatically split large uncompressed CSV files for parallel loading, so you usually don't need to manually split uncompressed CSV files. In certain cases where automatic file splitting isn't feasible due to data characteristics, manually splitting large CSVs might still benefit performance.

What is the file splitting guidance for the COPY command loading compressed CSV files?

The following table outlines the number of files you should use. When you reach the recommended number of files, you get better performance with larger files. The number of files is determined by the number of compute nodes multiplied by 60. For example, at 6000 DWU, you have 12 compute nodes, so you have 12 * 60 = 720 partitions. For a simple file splitting experience, see How to maximize COPY load throughput with file splits.

DWU	Number of files
100	60
200	60
300	60
400	60
500	60
1,000	120
1,500	180
2,000	240
2,500	300
3,000	360
5,000	600
6,000	720
7,500	900
10,000	1200
15,000	1800
30,000	3600

What is the file splitting guidance for the COPY command loading Parquet or ORC files?

You don't need to split Parquet and ORC files because the COPY command automatically splits files. For best performance, Parquet and ORC files in the Azure storage account should be 256 MB or larger.

Are there any limitations on the number or size of files?

There are no limitations on the number or size of files. However, for best performance, use files that are at least 4 MB. Also, limit the count of source files to a maximum of 5,000 files for better performance.

Are there any known issues with the COPY statement?

If you have an Azure Synapse workspace that you created before December 7, 2020, you might run into a similar error message when authenticating by using Managed Identity: com.microsoft.sqlserver.jdbc.SQLServerException: Managed Service Identity isn't enabled on this server. Please enable Managed Service Identity and try again.

To work around this problem, re-register the workspace's managed identity:

1. Install Azure PowerShell. See Install PowerShell.
2. Register your workspace's managed identity by using PowerShell:

   Connect-AzAccount
   Select-AzSubscription -SubscriptionId <subscriptionId>
   Set-AzSqlServer -ResourceGroupName your-database-server-resourceGroup -ServerName your-SQL-servername -AssignIdentity
   

Related content

• Loading overview with Azure Synapse Analytics