A shuffle format is available that does not preserve data distribution when the column values are not unique and also when it is conditionally masked. For example, consider the Original Table (Table 19-1) that shows two columns: EmpName and Salary. The Salary column has three distinct values: 10, 90, and 20.

Table 19-1 Original Table (Non-preservation)

EmpName | Salary |
---|---|

A |
10 |

B |
90 |

C |
10 |

D |
10 |

E |
90 |

F |
20 |

If you mask the Salary column with this format, each of the original values is replaced with one of the values from this set. Assume that the shuffle format replaces 10 with 20, 90 with 10, and 20 with 90 (Table 19-2).

Table 19-2 Mapping Table (Non-preservation)

EmpName | Salary |
---|---|

10 |
20 |

90 |
10 |

20 |
90 |

The result is a shuffled Salary column as shown in the Masked Table (Table 19-3), but the data distribution is changed. While the value 10 occurs three times in the Salary column of the Original Table, it occurs only twice in the Masked Table.

Table 19-3 Masked Table (Non-preservation)

EmpName | Salary |
---|---|

A |
20 |

B |
10 |

C |
20 |

D |
20 |

E |
10 |

F |
90 |

If the salary values had been unique, the format would have maintained data distribution.