When Trifacta® creates an imported dataset from a schematized source, the product applies its own type inferencing to the columns of the imported data. Type inferencing may be reapplied during some operations, such as the creation of samples or when data reshaping transformations are applied in the Transformer page.
If preferred, you can disable this type inferencing on the columns of your imported dataset. When the data is imported, the original types from the source system remain. Any types that do not have a corresponding match with the Trifacta data types must be manually typed in the application. You can use the following steps to disable type inference applied to a specific file during the import process.
Tip: For imported datasets from relational sources, you can identify in Flow View whether type inferencing has been applied to the dataset. When the dataset is selected in Flow View, locate the Type Inference entry in the right panel.
Tip: If you have already imported the dataset and need to change this setting, you can re-import the source and change the settings. In any flows that use the previously imported version of this dataset, you can change the input for any recipe that uses the old version to use this newly imported version through Flow View.
- After you have selected or specified the relational table to import in the Import Data page, click Edit Settings for the dataset card in the right panel.
- Deselect the Infer column data types checkbox.
- Continue the import process.
- When the dataset is loaded into the Transformer page, no new data typing is applied at all, unless you manually specify the Trifacta data types for the column.
This page has no comments.