Applies to: SQL Server Azure SQL Database Azure SQL Managed Instance Azure Synapse Analytics Analytics Platform System (PDW)
For SQL Server 2014 and lower, see Sqlcmd utility.
For using sqlcmd on Linux, see Install sqlcmd and bcp on Linux.
The sqlcmd utility lets you enter Transact-SQL statements, system procedures, and script files through a variety of available modes:
- At the command prompt.
- In Query Editor in SQLCMD mode.
- In a Windows script file.
- In an operating system (Cmd.exe) job step of a SQL Server Agent job.
The utility uses ODBC to execute Transact-SQL batches.
Download and install sqlcmd
Download Microsoft Command Line Utilities 15 for SQL Server (x64) (2.6 MB)
Download Microsoft Command Line Utilities 15 for SQL Server (x86) (2.3 MB)
The command line tools are General Availability (GA), however they're being released with the installer package for SQL Server 2019 (15.x).
- Release number: 15.0.2
- Build number: 15.0.2000.5
- Release date: September 11, 2020
The new version of SQLCMD supports Azure AD authentication, including Multi-Factor Authentication (MFA) support for SQL Database, Azure Synapse Analytics, and Always Encrypted features. The new BCP supports Azure AD authentication, including Multi-Factor Authentication (MFA) support for SQL Database and Azure Synapse Analytics.
System Requirements windows 11, Windows 10, Windows 7, Windows 8, Windows 8.1, Windows Server 2008 - 2022.
This component requires both the built-in Windows Installer 5 and the Microsoft ODBC Driver 17 for SQL Server.
See Install sqlcmd and bcp on Linux for instructions to install sqlcmd on Linux and macOS.
To check the SQLCMD version execute
sqlcmd -? command and confirm that 15.0.2000.5 version or higher is in use.
You need version 13.1 or higher to support Always Encrypted (
-g) and Azure Active Directory authentication (
-G). You may have several versions of sqlcmd.exe installed on your computer. Be sure you are using the correct version. To determine the version, execute
Azure Cloud Shell
You can try the sqlcmd utility from Azure Cloud Shell as it is pre-installed by default: Launch Cloud Shell
Azure Data Studio
To run sqlcmd statements in Azure Data Studio, select "Enable SQLCMD" from the editor toolbar.
SQL Server Management Studio (SSMS)
To run sqlcmd statements in SSMS, select SQLCMD Mode from the top navigation Query Menu dropdown.
SQL Server Management Studio (SSMS) uses the Microsoft .NET Framework SqlClient for execution in regular and SQLCMD mode in Query Editor. When sqlcmd is run from the command-line, sqlcmd uses the ODBC driver. Because different default options may apply, you might see different behavior when you execute the same query in SQL Server Management Studio in SQLCMD Mode and in the sqlcmd utility.
Currently, sqlcmd doesn't require a space between the command-line option and the value. However, in a future release, a space may be required between the command-line option and the value.
For more in-depth information on sqlcmd syntax and use, see:
sqlcmd -a packet_size -A (dedicated administrator connection) -b (terminate batch job if there is an error) -c batch_terminator -C (trust the server certificate) -d db_name -D -e (echo input) -E (use trusted connection) -f codepage | i:codepage[,o:codepage] | o:codepage[,i:codepage] -g (enable column encryption) -G (use Azure Active Directory for authentication) -h rows_per_header -H workstation_name -i input_file -I (enable quoted identifiers) -j (Print raw error messages) -k[1 | 2] (remove or replace control characters) -K application_intent -l login_timeout -L[c] (list servers, optional clean output) -m error_level -M multisubnet_failover -N (encrypt connection) -o output_file -p (print statistics, optional colon format) -P password -q "cmdline query" -Q "cmdline query" (and exit) -r[0 | 1] (msgs to stderr) -R (use client regional settings) -s col_separator -S [protocol:]server[instance_name][,port] -t query_timeout -u (unicode output file) -U login_id -v var = "value" -V error_severity_level -w screen_width -W (remove trailing spaces) -x (disable variable substitution) -X (disable commands, startup script, environment variables, optional exit) -y variable_length_type_display_width -Y fixed_length_type_display_width -z new_password -Z new_password (and exit) -? (usage)
Signs in to SQL Server with a Dedicated Administrator Connection (DAC). This kind of connection is used to troubleshoot a server. This connection works only with server computers that support DAC. If DAC isn't available, sqlcmd generates an error message, and then exits. For more information about DAC, see Diagnostic Connection for Database Administrators. The -A option isn't supported with the -G option. When connecting to SQL Database using -A, you must be a SQL server administrator. DAC isn't available for an Azure Active Directory administrator.
This switch is used by the client to configure it to implicitly trust the server certificate without validation. This option is equivalent to the ADO.NET option
TRUSTSERVERCERTIFICATE = true.
USE db_name statement when you start sqlcmd. This option sets the sqlcmd scripting variable SQLCMDDBNAME. This parameter specifies the initial database. The default is your login's default-database property. If the database doesn't exist, an error message is generated and sqlcmd exits.
Interprets the server name provided to -S as a DSN instead of a hostname. For more information, see DSN Support in sqlcmd and bcp in Connecting with sqlcmd.
The -D option is only available on Linux and MacOS clients. On Windows clients, it previously referred to a now-obsolete option which has been removed and is ignored.
Specifies the number of seconds before a sqlcmd login to the ODBC driver times out when you try to connect to a server. This option sets the sqlcmd scripting variable SQLCMDLOGINTIMEOUT. The default time-out for login to sqlcmd is eight seconds. When using the -G option to connect to SQL Database or Azure Synapse Analytics and authenticate using Azure Active Directory, a timeout value of at least 30 seconds is recommended. The login time-out must be a number between 0 and 65534. If the value supplied isn't numeric or doesn't fall into that range, sqlcmd generates an error message. A value of 0 specifies time-out to be infinite.
Uses a trusted connection instead of using a user name and password to sign in to SQL Server. By default, without -E specified, sqlcmd uses the trusted connection option.
The -E option ignores possible user name and password environment variable settings such as SQLCMDPASSWORD. If the -E option is used together with the -U option or the -P option, an error message is generated.
Sets the Column Encryption Setting to
Enabled. For more information, see Always Encrypted. Only master keys stored in Windows Certificate Store are supported. The -g switch requires at least sqlcmd version 13.1. To determine your version, execute
This switch is used by the client when connecting to SQL Database or Azure Synapse Analytics to specify that the user be authenticated using Azure Active Directory authentication. This option sets the sqlcmd scripting variable SQLCMDUSEAAD = true. The -G switch requires at least sqlcmd version 13.1. To determine your version, execute
sqlcmd -?. For more information, see Connecting to SQL Database or Azure Synapse Analytics By Using Azure Active Directory Authentication. The -A option isn't supported with the -G option.
-G option only applies to Azure SQL Database and Azure Synapse Analytics.
Azure AD Interactive Authentication is not currently supported on Linux or macOS. Azure AD Integrated Authentication requires Microsoft ODBC Driver 17 for SQL Server version 17.6.1 or higher and a properly Configured Kerberos environment.
Azure Active Directory Username and Password:
When you want to use an Azure Active Directory user name and password, you can provide the -G option and also use the user name and password by providing the -U and -P options.
Sqlcmd -S testsrv.database.windows.net -d Target_DB_or_DW -U email@example.com -P MyAADPassword -G
The -G parameter generates the following connection string in the backend:
SERVER = Target_DB_or_DW.testsrv.database.windows.net;UID= firstname.lastname@example.org;PWD=MyAADPassword;AUTHENTICATION = ActiveDirectoryPassword
Azure Active Directory Integrated
For Azure Active Directory Integrated authentication, provide the -G option without a user name or password. AAD Integrated Authentication requires Microsoft ODBC Driver 17 for SQL Server version 17.6.1 or higher and a properly configured Kerberos environment..
Sqlcmd -S Target_DB_or_DW.testsrv.database.windows.net -G
This generates the following connection string in the backend:
SERVER = Target_DB_or_DW.testsrv.database.windows.net Authentication = ActiveDirectoryIntegrated; Trusted_Connection=NO
-E option (Trusted_Connection) cannot be used along with the
Azure Active Directory Interactive
The Azure AD Interactive authentication for Azure SQL Database and Azure Synapse Analytics, allows you to use an interactive method supporting multi-factor authentication. For more information, see Active Directory Interactive Authentication.
Azure AD interactive requires sqlcmd version 15.0.1000.34 or later as well as ODBC version 17.2 or later.
To enable interactive authentication, provide -G option with user name (-U) only, without a password.
The following example exports data using Azure AD interactive mode indicating username where user represents an Azure AD account. This is the same example used in the previous section: Azure Active Directory Username and Password.
Interactive mode requires a password to be manually entered, or for accounts with multi-factor authentication enabled, complete your configured MFA authentication method.
sqlcmd -S testsrv.database.windows.net -d Target_DB_or_DW -G -U email@example.com
The previous command generates the following connection string in the backend:
SERVER = Target_DB_or_DW.testsrv.database.windows.net;UIDfirstname.lastname@example.org; AUTHENTICATION = ActiveDirectoryInteractive
In case an Azure AD user is a domain federated user using a Windows account, the user name required in the command-line, contains its domain account (for example, email@example.com see below):
sqlcmd -S testsrv.database.windows.net -d Target_DB_or_DW -G -U firstname.lastname@example.org
If guest users exist in a specific Azure AD and are part of a group that exists in SQL Database that has database permissions to execute the sqlcmd command, their guest user alias is used (for example, email@example.com).
There is a known issue when using the
-Uoption with SQLCMD, where putting the
-Uoption before the
-Goption may cause authentication to fail. Always start with the
-Goption followed by the
A workstation name. This option sets the sqlcmd scripting variable SQLCMDWORKSTATION. The workstation name is listed in the hostname column of the
sys.sysprocesses catalog view and can be returned using the stored procedure
sp_who. If this option isn't specified, the default is the current computer name. This name can be used to identify different
-j Prints raw error messages to the screen.
Declares the application workload type when connecting to a server. The only currently supported value is ReadOnly. If -K isn't specified, the sqlcmd utility won't support connectivity to a secondary replica in an Always On availability group. For more information, see Active Secondaries: Readable Secondary Replica (Always On Availability Groups)
Always specify -M when connecting to the availability group listener of a SQL Server availability group or a SQL Server Failover Cluster Instance. -M provides for faster detection of and connection to the (currently) active server. If -M isn't specified, -M is off. For more information about Listeners, Client Connectivity, Application Failover, Creation and Configuration of Availability Groups (SQL Server), Failover Clustering and Always On Availability Groups (SQL Server), and Active Secondaries: Readable Secondary Replicas(Always On Availability Groups).
This switch is used by the client to request an encrypted connection.
Is a user-specified password. Passwords are case-sensitive. If the -U option is used and the -P option isn't used, and the SQLCMDPASSWORD environment variable hasn't been set, sqlcmd prompts the user for a password. We don't recommend the use of the null password, but you can specify the null password by using a pair of contiguous double-quotation marks for the parameter value:
The use of -P should be considered insecure. Avoid giving the password on the command line. Alternatively, use the SQLCMDPASSWORD environment variable or interactively input the password by omitting the -P option.
We recommend that you use a strong password.
The password prompt is displayed by printing the password prompt to the console, as follows:
User input is hidden. This means that nothing is displayed and the cursor stays in position.
The SQLCMDPASSWORD environment variable lets you set a default password for the current session. Therefore, passwords don't have to be hard-coded into batch files.
The following example first sets the SQLCMDPASSWORD variable at the command prompt and then accesses the sqlcmd utility. At the command prompt, type:
SET SQLCMDPASSWORD= p@a$$w0rd
At the following command prompt, type:
If the user name and password combination is incorrect, an error message is generated.
The OSQLPASSWORD environment variable was kept for backward compatibility. The SQLCMDPASSWORD environment variable takes precedence over the OSQLPASSWORD environment variable. Now that OSQLPASSWORD is no longer shared, the utilities sqlcmd and osql can be used next to each other without interference. Old scripts will continue to work.
If the -P option is used with the -E option, an error message is generated.
If the -P option is followed by more than one argument, an error message is generated and the program exits.
Specifies the instance of SQL Server to which to connect. It sets the sqlcmd scripting variable SQLCMDSERVER.
Specify server_name to connect to the default instance of SQL Server on that server computer. Specify server_name [ \instance_name ] to connect to a named instance of SQL Server on that server computer. If no server computer is specified, sqlcmd connects to the default instance of SQL Server on the local computer. This option is required when you execute sqlcmd from a remote computer on the network.
protocol can be tcp (TCP/IP), lpc (shared memory), or np (named pipes).
If you don't specify a server_name [ \instance_name ] when you start sqlcmd, SQL Server checks for and uses the SQLCMDSERVER environment variable.
The OSQLSERVER environment variable has been kept for backward compatibility. The SQLCMDSERVER environment variable takes precedence over the OSQLSERVER environment variable; this means that sqlcmd and osql can be used next to each other without interference and that old scripts will continue to work.
Is the login name or contained database user name. For contained database users, you must provide the database name option (-d).
The OSQLUSER environment variable is available for backward compatibility. The SQLCMDUSER environment variable takes precedence over the OSQLUSER environment variable. This means that sqlcmd and osql can be used next to each other without interference. It also means that existing osql scripts will continue to work.
If neither the -U option or the -P option is specified, sqlcmd tries to connect by using Microsoft Windows Authentication mode. Authentication is based on the Windows account of the user who is running sqlcmd.
If the -U option is used with the -E option (described later in this article), an error message is generated. If the -U option is followed by more than one argument, an error message is generated and the program exits.
sqlcmd -U someuser -P s0mep@ssword -z a_new_p@a$$w0rd
Change password and exit:
sqlcmd -U someuser -P s0mep@ssword -Z a_new_p@a$$w0rd
-f codepage | i:codepage[,o:codepage] | o:codepage[,i:codepage]
Specifies the input and output code pages. The codepage number is a numeric value that specifies an installed Windows code page.
Code-page conversion rules:
If no code pages are specified, sqlcmd will use the current code page for both input and output files, unless the input file is a Unicode file, in which case no conversion is required.
sqlcmd automatically recognizes both big-endian and little-endian Unicode input files. If the -u option has been specified, the output will always be little-endian Unicode.
If no output file is specified, the output code page will be the console code page. This approach enables the output to be displayed correctly on the console.
Multiple input files are assumed to be of the same code page. Unicode and non-Unicode input files can be mixed.
Enter chcp at the command prompt to verify the code page of Cmd.exe.
Identifies the file that contains a batch of SQL statements or stored procedures. Multiple files may be specified that will be read and processed in order. Don't use any spaces between file names. sqlcmd will first check to see whether all the specified files exist. If one or more files don't exist, sqlcmd will exit. The -i and the -Q/-q options are mutually exclusive.
-i C:\<filename> -i \\<Server>\<Share$>\<filename> -i "C:\Some Folder\<file name>"
File paths that contain spaces must be enclosed in quotation marks.
This option may be used more than once: -iinput_file -II input_file.
Identifies the file that receives output from sqlcmd.
If -u is specified, the output_file is stored in Unicode format. If the file name isn't valid, an error message is generated, and sqlcmd exits. sqlcmd doesn't support concurrent writing of multiple sqlcmd processes to the same file. The file output will be corrupted or incorrect. See the -f switch is also relevant to file formats. This file will be created if it doesn't exist. A file of the same name from a prior sqlcmd session will be overwritten. The file specified here isn't the stdout file. If a stdout file is specified, this file won't be used.
-o C:< filename> -o \\<Server>\<Share$>\<filename> -o "C:\Some Folder\<file name>"
File paths that contain spaces must be enclosed in quotation marks.
-r[0 | 1]
Redirects the error message output to the screen (stderr). If you don't specify a parameter or if you specify 0, only error messages that has a severity level of 11 or higher are redirected. If you specify 1, all error message output including PRINT is redirected. Has no effect if you use -o. By default, messages are sent to stdout.
Causes sqlcmd to localize numeric, currency, date, and time columns retrieved from SQL Server based on the client's locale. By default, these columns are displayed using the server's regional settings.
Specifies that output_file is stored in Unicode format, regardless of the format of input_file.
Query Execution Options
Writes input scripts to the standard output device (stdout).
Sets the SET QUOTED_IDENTIFIER connection option to ON. By default, it's set to OFF. For more information, see SET QUOTED_IDENTIFIER (Transact-SQL).
-q " cmdline query "
Executes a query when sqlcmd starts, but doesn't exit sqlcmd when the query has finished running. Multiple-semicolon-delimited queries can be executed. Use quotation marks around the query, as shown in the following example.
At the command prompt, type:
sqlcmd -d AdventureWorks2012 -q "SELECT FirstName, LastName FROM Person.Person WHERE LastName LIKE 'Whi%';"
sqlcmd -d AdventureWorks2012 -q "SELECT TOP 5 FirstName FROM Person.Person;SELECT TOP 5 LastName FROM Person.Person;"
Do not use the GO terminator in the query.
If -b is specified together with this option, sqlcmd exits on error. -b is described later in this article.
-Q " cmdline query "
Executes a query when sqlcmd starts and then immediately exits sqlcmd. Multiple-semicolon-delimited queries can be executed.
Use quotation marks around the query, as shown in the following example.
At the command prompt, type:
sqlcmd -d AdventureWorks2012 -Q "SELECT FirstName, LastName FROM Person.Person WHERE LastName LIKE 'Whi%';"
sqlcmd -d AdventureWorks2012 -Q "SELECT TOP 5 FirstName FROM Person.Person;SELECT TOP 5 LastName FROM Person.Person;"
Do not use the GO terminator in the query.
If -b is specified together with this option, sqlcmd exits on error. -b is described later in this article.
Specifies the number of seconds before a command (or SQL statement) times out. This option sets the sqlcmd scripting variable SQLCMDSTATTIMEOUT. If a time_out value is not specified, the command does not time out. The query time_out must be a number between 1 and 65534. If the value supplied is not numeric or does not fall into that range, sqlcmd generates an error message.
The actual time out value may vary from the specified time_out value by several seconds.
-v var = value[ var = value...]
Creates a sqlcmdscripting variable that can be used in a sqlcmd script. Enclose the value in quotation marks if the value contains spaces. You can specify multiple var="values" values. If there are errors in any of the values specified, sqlcmd generates an error message and then exits.
sqlcmd -v MyVar1=something MyVar2="some thing"
sqlcmd -v MyVar1=something -v MyVar2="some thing"
Causes sqlcmd to ignore scripting variables. This parameter is useful when a script contains many INSERT statements that may contain strings that have the same format as regular variables, such as $(variable_name).
Specifies the number of rows to print between the column headings. The default is to print headings one time for each set of query results. This option sets the sqlcmd scripting variable SQLCMDHEADERS. Use -1 to specify that headers not be printed. Any value that is not valid causes sqlcmd to generate an error message and then exit.
-k [1 | 2]
Removes all control characters, such as tabs and new line characters from the output. This parameter preserves column formatting when data is returned. If 1 is specified, the control characters are replaced by a single space. If 2 is specified, consecutive control characters are replaced by a single space. -k is the same as -k1.
Specifies the column-separator character. The default is a blank space. This option sets the sqlcmd scripting variable SQLCMDCOLSEP. To use characters that have special meaning to the operating system such as the ampersand (&), or semicolon (;), enclose the character in quotation marks ("). The column separator can be any 8-bit character.
Specifies the screen width for output. This option sets the sqlcmd scripting variable SQLCMDCOLWIDTH. The column width must be a number greater than 8 and less than 65536. If the specified column width does not fall into that range, sqlcmd generates an error message. The default width is 80 characters. When an output line exceeds the specified column width, it wraps on to the next line.
This option removes trailing spaces from a column. Use this option together with the -s option when preparing data that is to be exported to another application. Cannot be used with the -y or -Y options.
Sets the sqlcmd scripting variable
SQLCMDMAXVARTYPEWIDTH. The default is 256. It limits the number of characters that are returned for the large variable length data types:
UDT (user-defined data types)
UDTs can be of fixed length depending on the implementation. If this length of a fixed length UDT is shorter that display_width, the value of the UDT returned is not affected. However, if the length is longer than display_width, the output is truncated.
Use the -y 0 option with extreme caution because it may cause serious performance issues on both the server and the network, depending on the size of data returned.
Sets the sqlcmd scripting variable
SQLCMDMAXFIXEDTYPEWIDTH. The default is 0 (unlimited). Limits the number of characters that are returned for the following data types:
char(n), where 1 <= n <= 8000
nchar(n), where 1 <= n <= 4000
varchar(n), where 1 <= n <= 8000
nvarchar(n), where 1 <= n <= 4000
varbinary(n), where 1 <= n <= 4000
Error Reporting Options
Specifies that sqlcmd exits and returns a DOS ERRORLEVEL value when an error occurs. The value that is returned to the DOS ERRORLEVEL variable is 1 when the SQL Server error message has a severity level greater than 10; otherwise, the value returned is 0. If the -V option has been set in addition to -b, sqlcmd will not report an error if the severity level is lower than the values set using -V. Command prompt batch files can test the value of ERRORLEVEL and handle the error appropriately. sqlcmd does not report errors for severity level 10 (informational messages).
If the sqlcmd script contains an incorrect comment, syntax error, or is missing a scripting variable, ERRORLEVEL returned is 1.
Controls which error messages are sent to stdout. Messages that has a severity level greater than or equal to this level are sent. When this value is set to -1, all messages including informational messages, are sent. Spaces are not allowed between the -m and -1. For example, -m-1 is valid, and -m -1 isn't.
This option also sets the sqlcmd scripting variable SQLCMDERRORLEVEL. This variable has a default of 0.
Controls the severity level that is used to set the ERRORLEVEL variable. Error messages that have severity levels greater than or equal to this value set ERRORLEVEL. Values that are less than 0 are reported as 0. Batch and CMD files can be used to test the value of the ERRORLEVEL variable.
Requests a packet of a different size. This option sets the sqlcmd scripting variable SQLCMDPACKETSIZE. packet_size must be a value between 512 and 32767. The default = 4096. A larger packet size can enhance performance for execution of scripts that have lots of SQL statements between GO commands. You can request a larger packet size. However, if the request is denied, sqlcmd uses the server default for packet size.
Specifies the batch terminator. By default, commands are terminated and sent to SQL Server by typing the word "GO" on a line by itself. When you reset the batch terminator, don't use Transact-SQL reserved keywords or characters that have special meaning to the operating system, even if they're preceded by a backslash.
Lists the locally configured server computers, and the names of the server computers that are broadcasting on the network. This parameter can't be used in combination with other parameters. The maximum number of server computers that can be listed is 3000. If the server list is truncated because of the size of the buffer a warning message is displayed.
Because of the nature of broadcasting on networks, sqlcmd may not receive a timely response from all servers. Therefore, the list of servers returned may vary for each invocation of this option.
If the optional parameter c is specified, the output appears without the Servers: header line, and each server line is listed without leading spaces. This presentation is referred to as clean output. Clean output improves the processing performance of scripting languages.
Prints performance statistics for every result set. The following display is an example of the format for performance statistics:
Network packet size (bytes): n
Clock Time (ms.): total t1 avg t2 (t3 xacts per sec.)
x = Number of transactions that are processed by SQL Server.
t1 = Total time for all transactions.
t2 = Average time for a single transaction.
t3 = Average number of transactions per second.
All times are in milliseconds.
If the optional parameter 1 is specified, the output format of the statistics is in colon-separated format that can be imported easily into a spreadsheet or processed by a script.
If the optional parameter is any value other than 1, an error is generated and sqlcmd exits.
Disables commands that might compromise system security when sqlcmd is executed from a batch file. The disabled commands are still recognized; sqlcmd issues a warning message and continues. If the optional parameter 1 is specified, sqlcmd generates an error message and then exits. The following commands are disabled when the -X option is used:
If the -X option is specified, it prevents environment variables from being passed on to sqlcmd. It also prevents the startup script specified by using the SQLCMDINI scripting variable from being executed. For more information about sqlcmd scripting variables, see Use sqlcmd with Scripting Variables.
Displays the version of sqlcmd and a syntax summary of sqlcmd options.
Options don't have to be used in the order shown in the syntax section.
When multiple results are returned, sqlcmd prints a blank line between each result set in a batch. In addition, the
<x> rows affected message doesn't appear when it doesn't apply to the statement executed.
To use sqlcmd interactively, type sqlcmd at the command prompt with any one or more of the options described earlier in this article. For more information, see Use the sqlcmd Utility
The options -L, -Q, -Z or -i cause sqlcmd to exit after execution.
The total length of the sqlcmd command-line in the command environment (Cmd.exe), including all arguments and expanded variables, is that which is determined by the operating system for Cmd.exe.
Variable Precedence (Low to High)
System-level environmental variables.
User-level environmental variables
Command shell (SET X=Y) set at command prompt before running sqlcmd.
:Setvar X Y
To view the environmental variables, in Control Panel, open System, and then select the Advanced tab.
sqlcmd Scripting Variables
|SQLCMDSTATTIMEOUT||-t||R/W||"0" = wait indefinitely|
|SQLCMDMAXFIXEDTYPEWIDTH||-Y||R/W||"0" = unlimited|
SQLCMDUSER, SQLCMDPASSWORD, and SQLCMDSERVER are set when :Connect is used.
R indicates the value can only be set one time during program initialization.
R/W indicates that the value can be modified by using the setvar command and subsequent commands will be influenced by the new value.
In addition to Transact-SQL statements within sqlcmd, the following commands are also available:
:XML [ON | OFF]
Be aware of the following when you use sqlcmd commands:
All sqlcmd commands, except GO, must be prefixed by a colon (:).
To maintain backward compatibility with existing osql scripts, some of the commands will be recognized without the colon, indicated by the [:].
sqlcmd commands are recognized only if they appear at the start of a line.
All sqlcmd commands are case insensitive.
Each command must be on a separate line. A command can't be followed by a Transact-SQL statement or another command.
Commands are executed immediately. They aren't put in the execution buffer as Transact-SQL statements are.
Starts the text editor. This editor can be used to edit the current Transact-SQL batch, or the last executed batch. To edit the last executed batch, the ED command must be typed immediately after the last batch has completed execution.
The text editor is defined by the SQLCMDEDITOR environment variable. The default editor is 'Edit'. To change the editor, set the SQLCMDEDITOR environment variable. For example, to set the editor to Microsoft Notepad, at the command prompt, type:
Clears the statement cache.
Prints the content of the statement cache.
:Setvar <var> [ "value" ]
Defines sqlcmd scripting variables. Scripting variables have the following format:
Variable names are case insensitive.
Scripting variables can be set in the following ways:
Implicitly using a command-line option. For example, the -l option sets the SQLCMDLOGINTIMEOUT sqlcmd variable.
Explicitly by using the :Setvar command.
By defining an environment variable before you run sqlcmd.
The -X option prevents environment variables from being passed on to sqlcmd.
If a variable defined by using :Setvar and an environment variable have the same name, the variable defined by using :Setvar takes precedence.
Variable names must not contain blank space characters.
Variable names can't have the same form as a variable expression, such as $(var).
If the string value of the scripting variable contains blank spaces, enclose the value in quotation marks. If a value for a scripting variable is not specified, the scripting variable is dropped.
Displays a list of the scripting variables that are currently set.
Only scripting variables that are set by sqlcmd, and those that are set using the :Setvar command will be displayed.
< filename >| STDERR|STDOUT
Redirect all error output to the file specified by file name, to stderr or to stdout. The Error command can appear multiple times in a script. By default, error output is sent to stderr.
Creates and opens a file that will receive the output. If the file already exists, it will be truncated to zero bytes. If the file is not available because of permissions or other reasons, the output will not be switched and will be sent to the last specified or default destination.
Switches error output to the stderr stream. If this has been redirected, the target to which the stream has been redirected will receive the error output.
Switches error output to the stdout stream. If this has been redirected, the target to which the stream has been redirected will receive the error output.
:Out < filename >| STDERR| STDOUT
Creates and redirects all query results to the file specified by file name, to stderr or to stdout. By default, output is sent to stdout. If the file already exists, it is truncated to zero bytes. The Out command can appear multiple times in a script.
:Perftrace < filename >| STDERR| STDOUT
Creates and redirects all performance trace information to the file specified by file name, to stderr or to stdout. By default performance trace output is sent to stdout. If the file already exists, it is truncated to zero bytes. The Perftrace command can appear multiple times in a script.
Execution Control Commands
:On Error[ exit | ignore]
Sets the action to be performed when an error occurs during script or batch execution.
When the exit option is used, sqlcmd exits with the appropriate error value.
When the ignore option is used, sqlcmd ignores the error and continues executing the batch or script. By default, an error message is printed.
Causes sqlcmd to exit.
[:] EXIT[ (statement) ]
Lets you use the result of a SELECT statement as the return value from sqlcmd. If numeric, the first column of the last result row is converted to a 4-byte integer (long). MS-DOS, Linux, and Mac pass the low byte to the parent process or operating system error level. Windows 200x passes the whole 4-byte integer. The syntax is:
You can also include the EXIT parameter as part of a batch file. For example, at the command prompt, type:
sqlcmd -Q "EXIT(SELECT COUNT(*) FROM '%1')"
The sqlcmd utility sends everything between the parentheses () to the server. If a system stored procedure selects a set and returns a value, only the selection is returned. The EXIT**()** statement with nothing between the parentheses executes everything before it in the batch and then exits without a return value.
When an incorrect query is specified, sqlcmd will exit without a return value.
Here is a list of EXIT formats:
Does not execute the batch, and then quits immediately and returns no value.
- :EXIT( )
Executes the batch, and then quits and returns no value.
Executes the batch that includes the query, and then quits after it returns the results of the query.
If RAISERROR is used within a sqlcmd script and a state of 127 is raised, sqlcmd will quit and return the message ID back to the client. For example:
RAISERROR(50001, 10, 127)
This error will cause the sqlcmd script to end and return the message ID 50001 to the client.
The return values -1 to -99 are reserved by SQL Server, and sqlcmd defines the following additional return values:
|-100||Error encountered prior to selecting return value.|
|-101||No rows found when selecting return value.|
|-102||Conversion error occurred when selecting return value.|
GO signals both the end of a batch and the execution of any cached Transact-SQL statements. The batch is executed multiple times as separate batches. You cannot declare a variable more than once in a single batch.
:r < filename >
Parses additional Transact-SQL statements and sqlcmd commands from the file specified by < filename > into the statement cache.
If the file contains Transact-SQL statements that are not followed by GO, you must enter GO on the line that follows :r.
< filename > is read relative to the startup directory in which sqlcmd was run.
The file will be read and executed after a batch terminator is encountered. You can issue multiple :r commands. The file may include any sqlcmd command. This includes the batch terminator GO.
The line count that is displayed in interactive mode will be increased by one for every
:r command encountered. The
:r command will appear in the output of the list command.
Lists the locally configured servers and the names of the servers broadcasting on the network.
:Connect server_name[\instance_name] [-l timeout] [-U user_name [-P password]]
Connects to an instance of SQL Server . Also closes the current connection.
|n>0||wait for n seconds|
The SQLCMDSERVER scripting variable will reflect the current active connection.
If timeout is not specified, the value of the SQLCMDLOGINTIMEOUT variable is the default.
If only user_name is specified (either as an option, or as an environment variable), the user will be prompted to enter a password. Users are not prompted if the SQLCMDUSER or SQLCMDPASSWORD environment variables have been set. If neither options nor environment variables are provided, Windows Authentication mode is used to sign in. For example to connect to an instance,
instance1, of SQL Server,
myserver, by using integrated security you would use the following command:
To connect to the default instance of
myserver using scripting variables, you would use the following:
:setvar myusername test
:setvar myservername myserver
:connect $(myservername) $(myusername)
[:] !!< command>
Executes operating system commands. To execute an operating system command, start a line with two exclamation marks (!!) followed by the operating system command. For example:
The command is executed on the computer on which sqlcmd is running.
:XML [ON | OFF]
For more information, see XML Output Format and JSON Output Format in this article
Lists sqlcmd commands together with a short description of each command.
sqlcmd File Names
sqlcmd input files can be specified with the -i option or the :r command. Output files can be specified with the -o option or the :Error, :Out and :Perftrace commands. The following are some guidelines for working with these files:
:Error, :Out and :Perftrace should use separate <filename>. If the same <filename> is used, inputs from the commands may be intermixed.
If an input file that is located on a remote server is called from sqlcmd on a local computer and the file contains a drive file path such as :Out c:\OutputFile.txt. The output file is created on the local computer and not on the remote server.
Valid file paths include:
"C:\Some Folder\<file name>". If there is a space in the path, use quotation marks.
Each new sqlcmd session will overwrite existing files that have the same names.
sqlcmd prints any informational message that is sent by the server. In the following example, after the Transact-SQL statements are executed, an informational message is printed.
At the command prompt, type the command:
At the sqlcmd prompt type:
When you press ENTER, the following informational message is printed: "Changed database context to 'AdventureWorks2012'."
Output Format from Transact-SQL Queries
sqlcmd first prints a column header that contains the column names specified in the select list. The column names are separated by using the SQLCMDCOLSEP character. By default, this is a space. If the column name is shorter than the column width, the output is padded with spaces up to the next column.
This line is followed by a separator line that is a series of dash characters. The following output shows an example.
Start sqlcmd. At the sqlcmd command prompt, type the query:
SELECT TOP (2) BusinessEntityID, FirstName, LastName
When you press ENTER, the following result set is returned.
BusinessEntityID FirstName LastName
---------------- ------------ ----------
285 Syed Abbas
293 Catherine Abel
(2 row(s) affected)
BusinessEntityID column is only four characters wide, it has been expanded to accommodate the longer column name. By default, output is terminated at 80 characters. This can be changed by using the -w option, or by setting the SQLCMDCOLWIDTH scripting variable.
XML Output Format
XML output that is the result of a FOR XML clause is output, unformatted, in a continuous stream.
When you expect XML output, use the following command:
sqlcmd returns error messages in the usual format. Notice that the error messages are also output in the XML text stream in XML format. By using
:XML ON, sqlcmd does not display informational messages.
To set the XML mode to off, use the following command:
The GO command should not appear before the XML OFF command is issued because the XML OFF command switches sqlcmd back to row-oriented output.
XML (streamed) data and rowset data can't be mixed. If the XML ON command hasn't been issued before a Transact-SQL statement that outputs XML streams is executed, the output is garbled. Once the XML ON command has been issued, you can't execute Transact-SQL statements that output regular row sets.
:XML command does not support the SET STATISTICS XML statement.
JSON Output Format
When you expect JSON output, use the following command:
:XML ON. Otherwise the output includes both the column name and the JSON text. This output is not valid JSON.
To set the XML mode to off, use the following command:
For more info, see XML Output Format in this article.
Use Azure Active Directory Authentication
Examples using Azure Active Directory Authentication:
sqlcmd -S Target_DB_or_DW.testsrv.database.windows.net -G -l 30 sqlcmd -S Target_DB_or_DW.testsrv.database.windows.net -G -U firstname.lastname@example.org -P MyAADPassword -l 30
sqlcmd Best Practices
Use the following practices to help maximize security and efficiency.
Use integrated security.
Use -X in automated environments.
Secure input and output files by using appropriate NTFS file system permissions.
To increase performance, do as much in one sqlcmd session as you can, instead of in a series of sessions.
Set time-out values for batch or query execution higher than you expect it will take to execute the batch or query.
Use the following practices to help maximize correctness:
Use -V16 to log any severity 16 level messages. Severity 16 messages indicates general errors that can be corrected by the user.
Check the exit code and DOS ERRORLEVEL variable after the process has exited. sqlcmd will return 0 normally, otherwise it will set the ERRORLEVEL as configured by -V. In other words, ERRORLEVEL shouldn't be expected to be the same value as the Error Number reported from SQL Server. The Error Number is a SQL Server-specific value corresponding to the system function @@ERROR. ERRORLEVEL is a SQLCMD-specific value to indicate why it (i.e. SQLCMD) terminated, and it's value is influenced by specifying -b command line argument.
Using -V16 in combination with checking the exit code and DOS ERRORLEVEL can help catch errors in automated environments, particularly quality gates before a production release.
Submit and view feedback for