Query data using MATLAB

  • Article

MATLAB is a programming and numeric computing platform used to analyze data, develop algorithms, and create models. This article explains how to get an authorization token in MATLAB for Azure Data Explorer, and how to use the token to interact with your cluster.

Prerequisites

Select the tab for the operating system used to run MATLAB.

  1. Install JDK 8 and Maven.

  2. Clone the MATLAB Interface for Azure Services repository:

    git clone https://github.com/mathworks-ref-arch/matlab-azure-services.git

  3. Build the MATLAB Azure SDK jar:

    cd matlab-azure-services/Software/Java
mvn clean package

    The build should produce the file: Software/MATLAB/lib/jar/azure-common-sdk-0.2.0.jar.

  4. In the MATLAB studio, load the jar file and add it to the java static class path:

    edit(fullfile(prefdir,'javaclasspath.txt'));

  5. In java static class path file, add an entry corresponding to the jar file:

    (full-path)/matlab-azure-services/Software/MATLAB/lib/jar/azure-common-sdk-0.2.0.jar

  6. Restart the MATLAB IDE. Once reloaded, run the startup.m script available at matlab-azure-services\Software\MATLAB\startup.m. This ensures all the prerequisite functions are set up for access to Azure services. The output should look as follows:

    >> startup
Adding Azure Paths
------------------
Adding: matlab-azure-services/Software/MATLAB/app
Adding: matlab-azure-services/Software/MATLAB/app/functions
Adding: matlab-azure-services/Software/MATLAB/app/system
Adding: matlab-azure-services/Software/MATLAB/lib
Adding: matlab-azure-services/Software/MATLAB/config
Skipping: matlab-azure-services/Software/Utilities
Checking the static Java classpath for: matlab-azure-services/Software/MATLAB/lib/jar/azure-common-sdk-0.2.0.jar
Found: azure-common-sdk-0.2.0.jar

Perform user authentication

With user authentication, the user is prompted to sign-in through a browser window. Upon successful sign-in, a user authorization token is granted. This section shows how to configure this interactive sign-in flow.

To perform user authentication:

  1. Create an Auth.json file in your working directory with the relevant credentials. The KustoClientAppId returned from https://<adx-cluster>.kusto.chinacloudapi.cn/v1/rest/auth/metadata needs to be replaced with the value of ClientId below:

    {
    "AuthMethod": "InteractiveBrowser",
    "TenantId" : "<AAD Tenant / Authority ID for the login>",
    "ClientId" : "<Client ID>",
    "RedirectUrl": "http://localhost:8675"
}

  2. Query your cluster as follows:

    % References the credential file created in the previous step
credentials = configureCredentials('Auth.json');

% Point the scopes to the Azure Data Explorer cluster that we want to query
request = azure.core.credential.TokenRequestContext();
request.addScopes('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/.default>');
token=credentials.getToken(request);

% Prepare to query the cluster
options=weboptions('HeaderFields',{'RequestMethod','POST';'Accept' 'application/json';'Authorization' ['Bearer ', token.getToken()]; 'Content-Type' 'application/json; charset=utf-8'; 'Connection' 'Keep-Alive'; 'x-ms-app' 'Matlab'; 'x-ms-client-request-id' 'Matlab-Query-Request'});

% The DB and KQL variables represent the database and query to execute
querydata = struct('db', "<DB>", 'csl', "<KQL>");
querryresults  = webwrite('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/v2/rest/query>', querydata, options);

% The results row can be extracted as follows
results=querryresults{3}.Rows

Perform application authentication

Microsoft Entra application authorization can be used for scenarios where interactive sign-in isn't desired and automated runs are necessary.

To perform application authentication:

  1. Create an Auth.json file in your working directory with the relevant credentials:

    {
    "AuthMethod": "ClientSecret",
    "TenantId" : "<AAD Tenant / Authority ID for the login>",
    "ClientId" : "<Client ID of the Azure AD application>",
    "ClientSecret": "<Client Key of the Azure AD application>",
    "RedirectUrl": "http://localhost:8675"
}

  2. Query your cluster as follows:

    % References the credential file created in the previous step
credentials = configureCredentials('Auth.json');

% Point the scopes to the Azure Data Explorer cluster that we want to query
request = azure.core.credential.TokenRequestContext();request.addScopes('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/.default>');
token=credentials.getToken(request);

% Prepare to query the cluster
options=weboptions('HeaderFields',{'RequestMethod','POST';'Accept' 'application/json';'Authorization' ['Bearer ', token.getToken()]; 'Content-Type' 'application/json; charset=utf-8'; 'Connection' 'Keep-Alive'; 'x-ms-app' 'Matlab'; 'x-ms-client-request-id' 'Matlab-Query-Request'});

% The DB and KQL variables represent the database and query to execute
querydata = struct('db', "<DB>", 'csl', "<KQL>");
querryresults  = webwrite('<https://adx-cluster-changeme.kusto.chinacloudapi.cn/v2/rest/query>', querydata, options);

% The results row can be extracted as follows
results=querryresults{3}.Rows

Related content