You can collect data from tables on web pages using the Capture Table action. This action recognizes the headers in a table and collects data from each column and row. If the Capture Table action captures several similar tables, it can recognize new columns on the fly and automatically collect them.
- In the Agent Builder, open an agent.
- Select Add Action > Capture Table.
- Click a header cell on a table within the web page.
- Follow the steps that will appear on the top left of the Agent Builder after you have clicked the first header.
- Click another header cell.
- Click two non-header cells in the same column to discover row.
- Click two cells in the same row to discover cells.
- Click SAVE.
The captured table data appears in the Captured Text Preview panel at the bottom of the Agent Builder.
Managing Actions in a Table Capture List
- Rename fields – By default, the agent names columns in the output collection after the columns in the table. You can change a column name without affecting the accuracy of the collected data by renaming it in the Capture table action.
- Refine Capture Text – Limit which tables rows you harvest.