Skip to main content

Dataset schema

Review and edit the column structure on a dataset's Schema tab. An object dataset does not have schema editing controls.

Edit the schema​

When you edit a schema, you change the column structure and order, then check the saved result.

  1. Select the dataset in the collection tree and open the Schema tab.
  2. Select Edit in the detail header.
  3. Add columns, or change a column's name, alias, description, type, or whether NULL is allowed.
  4. Drag rows or use the arrow keys to reorder columns.
  5. Select Save Changes and confirm that the new schema appears.

For a rest dataset, also specify a JSONPath for each column. For a column that already contains loaded data, you cannot change its name, type, or whether NULL is allowed; you can edit only its alias and description.

View options​

  • Table: Shows each column's name, alias, description, data type, and whether NULL is allowed, one row per column.
  • JSON: Shows the raw schema definition.
  • Show native types: Shows data types by their Arrow type names.

In the data type picker, browse types by group or search for them by name.

Column naming rules​

A column name must begin with an English letter or an underscore (_). After that, it can contain only English letters, numbers, and underscores.

  • Empty names, Korean characters, spaces, /, \, and other symbols are not allowed.
  • SQL reserved words of the analytical engine are not allowed.
  • A schema with duplicate column names cannot be saved.
  • The maximum length is 128 characters.

Enter a display name as the alias.

Supported data types​

CategoryExample types
String and binarystring, large_string, binary, large_binary
Integerint8, int16, int32, int64, uint8, uint16, uint32, uint64
Floating point and fixed pointfloat16, float32, float64, decimal128, decimal256
Date and timedate32, date64, timestamp, time32, time64, duration, interval
Otherbool, list, struct, map, dictionary, null

Temporary schema​

You can create a new delta dataset with a temporary schema that contains only field1. When you upload the first CSV or Parquet file, the schema is initialized from the file columns.

note

Dataset schemas have no option to set a primary key or display column. Manage these settings in ontology entities and relationships.

Next steps​

  • Upload files or fill sample data in Dataset data.
  • Review the creation fields for each type and the detail tabs in Datasets.