Skip to main content

Remove Initial Structure

When you import a dataset from a file, the Designer Cloud Powered by Trifacta platform attempts to detect the structure of the file and to apply an initial set of parsing steps to the data to render it in tabular form for display in the Transformer page. For example, JSON files may be turned into a table of data as long as the structure of the data supports this structuring.


Initial parsing steps are applied only to file-based sources of data.

These steps vary based on the file format of data that is being imported. Depending on the dataset, you may need to modify these steps or rebuild them altogether. You can use the following steps to prevent theDesigner Cloud Powered by Trifacta platform from detecting the structure and automatically hiding these steps.


You should allow the product to detect the structure first. If it does not detect the structure well, you can experiment with disabling it and rebuilding the steps to render your data in tabular format.


If you have already imported the dataset and need to change this setting, you can re-import the source and change the settings. Wherever the previously imported version of this dataset is referenced, you must change the reference to use this newly imported dataset.


When the steps are completed, the initial parsing steps are listed in any recipe that you create from the imported dataset. If you wish to remove them altogether, you can delete them from the recipe.

Import Unstructured Dataset

  1. After you have selected or specified the file to import in the Import Data page, click Edit Settings for the dataset card in the right panel.

  2. Deselect the Detect Structure checkbox. For more information, see File Import Settings.

  3. Continue the import process by adding the dataset.

  4. When the imported dataset is added, it is listed as an unstructured dataset.

Use Unstructured Dataset

  1. Import the unstructured dataset in Flow View.

  2. Select the unstructured dataset and click Add new recipe.

  3. When you select the recipe, the initial parsing steps are listed in the right panel.

  4. When the dataset is loaded into the Transformer page, you can modify these steps to improve the parsing or delete them altogether.


    Any step that breaks up the data into individual rows into individual rows must be the first step in the recipe. To create, enter Break into rows in the Search panel.