SharePoint Online

Cloud Connector reads Excel workbooks, CSV, JSON, XML and text files from SharePoint document libraries through Microsoft Graph, as the account that signs in.

Setting up

1. An administrator adds the delegated permission Sites.Read.All to the deployment's Cloud Connector app registration and grants admin consent, and makes a service account a member of each site it should read.
2. Install the SharePoint Online — files template from the Templates gallery. It creates a single-instance source system with one instance, Service account, set to Microsoft sign-in.
3. Open the instance, Credentials, and press Connect…. Sign in as the service account. The loads run with that account's permissions.

Only a source system installed from the template shows the SharePoint catalogue; a Graph source built by hand is a plain cloud source.

The catalogue

Pick a Site from the list or paste its URL (a site granted recently can take hours to reach search). The site's libraries are read whole and shown as a tree, pruned to folders that lead to a loadable file; libraries with nothing loadable are hidden.

A file row shows its format and folder path, the last editor and date, Preview and Open in SharePoint. A workbook expands into sheet rows, and each sheet into its table rows:

• A sheet is the whole worksheet; its first row must hold the column names.
• A table is a named Excel table (Insert → Table). It reads only its own range with its own header row, so one sheet can feed several datasets.

Preview is the way to assign: it creates an endpoint for the file (by id, so it survives renames and moves), opens the standard preview with Sheet: / Table: chips to switch, and Create dataset opens the usual assign dialog. The dataset is named after the table, else the sheet, else the file. An assigned file, sheet or table leaves the tree and returns when its dataset is deleted. Re-read from SharePoint refreshes the tree.

.xlsx and .xlsm are read as Excel (.xls is not), .csv and .txt as CSV, plus .json and .xml.