Register Sources and Scan Metadata. Purview supports registering a number of Azure based and non-Azure based sources. These sources are added to collections which are used to organize assets and sources by our businesses' flow. They're also the tool used to manage access across Microsoft Purview. Here are some of the data sources currently available in Purview. To register a source, select register from the data map, then select an available source. 

This will add the source to a selected collection. After data sources are registered in our Microsoft Purview account, the next step is to scan the data sources. The scanning process establishes a connection to the data source and captures technical metadata like names, file size, columns, and so on. It also extracts schema for structured data sources, applies classifications on schemas, and applies sensitivity labels if our Purview data map is connected to a Purview compliance portal. The scanning process can be triggered to run immediately or can be scheduled to run on a periodic basis to keep our Purview account up to date. Microsoft Purview is secure by default. 

No passwords or secrets are stored directly in the program, so we'll need to choose an authentication method for our sources. There are a few possible ways to authenticate our Purview account, but not all methods are supported for each data source. These authentication pathways include Managed Identity, Service Principal, SQL Authentication, Account Key, or Basic Authentication. When scanning a source, we have a choice to scan the entire data source or choose only specific entities like folders or tables to scan. Available options depend on the source we're scanning and can be defined for both one-time and scheduled scans. A scan rule set determines the kinds of information a scan will look for  when it's  running against one of our sources. 

Available rules depend on the kind of source we're scanning but include things like the file types we should scan and the kinds of classifications we need. There are over 200 built-in system classifications that can be selected as well as the ability to create custom classifications. For a catalog to know if a file, table, or container was deleted, it compares the last scan output against the current file scan output. For example, suppose that the last time we scanned an Azure Data Lake storage gen2 account, it included a folder named Folder 1. When the same account is scanned again, Folder 1 is missing. Therefore, the catalog assumes the folder has been deleted.