Example: Mask Tables With Linked Seed Data (Teradata)

This scenario shows how you mask multiple columns in a table whose seed data relates to each other. The use case for this masking scenario is that you are masking, for example, an address table where the city, state, and zip code need to match. This example uses a Teradata database.
tdm10
This scenario shows how you mask multiple columns in a table whose seed data relates to each other. The use case for this masking scenario is that you are masking, for example, an address table where the city, state, and zip code need to match. This example uses a Teradata database.
You want to ensure that the seed data has matching data columns, and that the hash operates on the same columns.
Follow these steps:
  1. Issue the following SQL to find which data categories have linked data. Substitute the database name where you have installed the masking objects for FUNCTIONS:
    SELECT DISTINCT rd_ref_id FROM
    FUNCTIONS.gtsrc_reference_data
    WHERE rd_ref_value2 IS NOT NULL
    The query returns a list of linked
    rd_ref_id
    s.
  2. Pick a linked data category and query it in the seed table. Pick specific columns in the result to use for each column mask. For example:
    SELECT * FROM
    FUNCTIONS.gtsrc_reference_data
    WHERE rd_ref_id = 'US ADDRESS'
  3. Create a transformation map of type TERADATA.
  4. Set the column masks according to the data.
    • Type the data category name if it does not exist in the drop down, but does exist in the seed data table.
      Example: Type
      HASHLOV,US ADDRESS
      in the Transformation column.
    • Ensure that the List Col#  number matches the corresponding column in the seed data table.
    • To link the columns in the table to be masked, ensure that the hash operates on the same column in the table, so that values are picked from the same row in the seed table.
  5. Specify the column by putting the column name in the notes section of the transformation map.
    In general, you use a primary key column, or at least a column with as many unique values as possible.
    Example:
    ADDRESS_ID
image2017-5-25 16:51:8.png