• replace last names with XXX[..]XXX.
• replace city names with 1, 2, 3, 4, 5, ... to compile statistics on the number of inhabitants without being able to identify the cities.

• static data masking, which is irreversible: data is permanently lost.
• dynamic data masking. This type of data masking is applied in near real-time by the database server: it is therefore reversible. Initial data is accessible depending the user privileges.

• Static data masking eliminates all personal data: the file therefore falls outside the scope of the GDPR. The file can be processed, stored and shared without any restrictions.
• In the case of dynamic data masking, personal data is not eliminated: the file therefore remains subject to the GDPR (e.g. with regard to retention periods).

Use conditions

Principle

• defining data masking rules ("masks"). A rule defines the data masking mode used and its characteristics. These rules can be implemented:
  • in the data model editor, for each desired item.
  • via SQL statements.
  Caution: You can only define one rule per item. It is not possible to combine several rules.
• defining data masking user privileges. Users can be allowed or prevented from viewing masked items. These privileges can be defined:
  • in the HFSQL Control Center.
  • programmatically, using HFSQL functions.
  • programmatically, using SQL statements.

Modifying the analysis

1. Select the desired file in the data model editor.
2. Right-click the data file and select "Description of items".
3. Select the item you want to mask.
4. Go to the "Data mask" tab.
   
   "No rule (the item is not masked)" is selected by default.
5. Select and configure the desired rule:
   • random text (lorem ipsum), to generate several random words,
   • a constant value (to be defined),
   • noise (a variation of plus or minus x %) for integer values,
   • a value returned by a stored procedure, for a specific data mask (for example, to mask a section of code while preserving part of the data),
   • partial scrambling, to keep some characters at the beginning and end of values, but replacing the rest with a given character.
6. Confirm changes in the description window of items.
7. Generate the analysis and run the automatic data modification.

• The MaskingRule property gets the data masking rule defined for a given item. This rule is in SQL format.
• To find out which items have been masked and which rule has been used, go the "Analysis" tab, "Analysis" group, expand "Maintenance" and select "List of items with a masking rule".

Enabling data file encryption

1. Select the data file.
2. In the context menu, select "Data file description".
3. Go to the "Details" tab, "Data protection" section and specify the desired encryption mode for the data file, index file and/or memo file.
   The encryption algorithm to be used by default is RC5 (16 rounds).

• for the database,
• for the data file,
• for the item.

• from the HFSQL Control Center,
• programmatically.

HFSQL Control Center

To grant permissions at the database or data file level, simply click "Manage rights" in the HFSQL Control Center ribbon.

To grant permissions at the item level, simply click the "lock" icon of the item in the "Description" tab.

In the window that appears, you can change the permissions set on the item.

Programming

// Allow "MARK" to unmask data:
// - all items from all files in the "DataMasking" database
HModifyDatabaseRights(ctMyConnection, "MARK", hRightsUnmask, hAllowed, "DataMasking")
// - all the items of the "Address"
HModifyFileRights(ctMyConnection, "MARK", hRightsUnmask, hAllowed, "DataMasking", "Address")
// - only the "PostalCode" item in the "Address" file
HModifyItemRights(ctMyConnection, "MARK", hRightsUnmask, hAllowed, "DataMasking", "Address", "PostalCode")

// Checks if the permission has been granted at the data file level
IF HInfoFileRights(gctConnection, sUser, hRightsUnmask, 
		DATABASE_NAME, Address, hEffective) = hForbidden THEN
	RETURN False
END

• ADD MASKED: Combined with the ALTER TABLE SQL statement, creates a data mask for an item in a data file.
• DROP MASKED: Combined with the ALTER TABLE SQL statement, deletes a data mask of an item in a data file.
• GRANT: Enables a data mask.
• REVOKE: Disables a data mask.

• Unmasking permissions are required to:
  • use HReadSeek on a masked key item.
  • use a SQL query with a WHERE condition on a masked item.
  • use an SQL query with a reference to a masked item in the ORDER BY clause.
  • use HBackup and HCopyFile with data files.
  • use data file replication.
• HCopyRecord function: original values cannot be retrieved. Masked values are buffered in the new record.
• You can apply data masks to key or non-key items. However, it is not possible to apply a mask to composite key items.
• If you apply a data mask to an item of type array, this mask applies to all the elements of the array.
• Materialized view:
  • Unmasking permissions are required to to define a materialized view on a file.
  • Even if unmasking permissions are granted, items remain masked in the materialized view.
  • Regardless of the masking/unmasking permissions, it is not possible to create a materialized view with masked calculated items.
• Even if unmasking permissions are granted, it is not possible to make backup copies of files containing masked items.

Adding and modifying records (HFSQL functions)

• When adding a new record with HAdd or HSave, items are initialized either with their default value, or with the specified value. Data masks are not applied.
• When modifying a record with HSave or HModify, only masked items on which the user has unmasking permissions will be modified.

INSERT and UPDATE SQL queries

• In an INSERT SQL query, items without values are initialized with the default value defined in the analysis. Items with values are automatically initialized without applying the data mask.
• In an UPDATE SQL query, items with values are automatically initialized without applying the data mask.

• If the item is masked and has values in the SQL query, the mask will not be applied to the returned value.
• If the item is masked and does not have values in the SQL query, the mask will be applied to the returned value.