Skip to main content Skip to complementary content

Exploring semantic categories of data columns

About this task

The example below uses a database table which holds customer information.


  1. In the DQ Repository tree view, expand Metadata and browse to the table you want to analyze.
  2. Right-click the table and select Semantic-aware Analysis, or right-click a set of columns in the table and select Semantic-aware Analysis.
    Contextual menu of a table under the Metadata node.

    The semantic wizard opens listing all the columns of the table or listing the selected set of columns depending on whether you started the analysis on a table or on a set of columns respectively. The Category line in the wizard assigns semantic categories for the matched columns.

    Overview of the Semantic Category Inference wizard.
  3. Configure the Sampling Options:
    • Sampling Strategy: define what to list in the data preview. Select First N Rows to list the N first data records or select Reservoir Sampling to list N random records. Then set the number of records in the Number of rows field.
    • Threshold for category discovery: decide the minimum threshold for the matches to show in the Category lists of the analyzed columns.

      This threshold filters the less probable categories of the analyzed columns.

    • Refresh: refresh the data preview after any change in the configuration.
  4. From the Category field of each of the matched columns, either:
    • Select a category of data from the Category list that best suites the column, or
    • Enter a meaningful name for the column that best represent the content.
  5. To edit the name of a column, click in the field twice, type the name and press Enter on your keyboard to save the changes.
    The names entered by you will display in a different color. This step stores locally the categories and the semantic names of the columns. If no semantic names are found, categories are stored anyway.
    This is not mandatory but will help you better match table metadata with the concepts stored in the ontology repository on the Elasticsearch server.

    The percentages of the proposed categories are calculated by analyzing the data in the columns against the following methods: regex, data dictionary and keyword dictionary. The dictionary indexes and regex categories are embedded in Talend Studio and are used to decide what category does the data fall in.

  6. Click Next to open a page in the wizard where you can see the results of matching column metadata and semantic concepts with the concepts in the ontology repository.

Did this page help you?

If you find any issues with this page or its content – a typo, a missing step, or a technical error – let us know how we can improve!