After you open the Table Wizard according to the instructions in Starting the Schema Tool, you can generate a normalized view of your data by taking the following steps:

  1. Select Normalized View under Select an initial view for your data in the Table Wizard window.

    The Table Wizard displays a graphic which shows how objects, such as embedded documents and arrays, are normalized into a relational structure.


    table wizard landing page

  2. Click OK to generate the normalized schema.
    The Schema Tool's main display appears with a normalized view of your data in the left-hand panel.
    Note: If you want to change the initial view of your data later, select File>Restart Wizard from the Table Wizard dialog. Be aware that restarting the wizard will also overwrite any customizations you have made to your schema map. See Restarting the Wizard for more information.
  3. Click the Save button to save the schema.
  4. Click database and table nodes to tour the normalized schema structure.

A normalized view of your data has been generated and saved.


main display with normalized schema and open nodes

The Table Wizard has decomposed each collection into a parent table with multiple child tables. In the parent table, the _id field is mapped as a primary key column and is denoted with the ID icon . The parent table also contains any columns derived from simple types (in contrast to subdocuments and arrays). In turn, child tables are derived from any complex types discovered in the source data. The child tables share a foreign key relationship with the parent table. When a subdocument is normalized into a child table, the primary key column and the foreign key column are one and the same. Such a column is denoted with the ID/FK icon . When an array is normalized into a child table, the primary key column and foreign key column are distinct columns. The primary key column is denoted with the ID icon , while the foreign key column is denoted with the FK icon . For more information on relational mapping, refer to "Mapping Objects to Tables" in your driver documentation.

Depending on your work flow, you can now proceed with customizing the schema. See Customizing Your Schema for details.