Transform

Goal

Use transform to parse the values created by the anonymization rule. The transform is a java function that takes the output of the format string as input, and processes them before storing the results to the database.

The difference between a convert and a transform is that a convert is done on an input source and a transform is done on the output of the anonymization rule, i.e., the output of the format string.

Use cases for transform

Use transform to ensure that the generated anonymized data follows business specific structural rules.

Ano Structure and Examples

Structure

...
        <rule> <column>
            format %s
            transform <transformation>
            <input-source>
...

• Passes the value of %s to the <transformation> function. The result is then stored database.

Custom Structure

transformation <package-location>.<transformation>

task <task-name> {
    ...
        <rule> <column>
            format %s
            transform <transformation>
            <input-source>
}

• Custom Transformation: If you use a custom transformation and not a built-in one, then import it in the .ano file using the transformation keyword
• import the custom function from its package in the java project

• `transformation example.anonymizer.transformations.PostCodeGeneralization

  `

Email transform Example

...
    update CUSTOMER
        mask EMAIL
            format "%s"
            transform Email
            unique
            column NAME
...

CreditCard transform Example

    update Customer
        mask CreditCard
            format "%d%d"
            transform CreditCard
            random-integer 40005000 60000000
            random-integer 10001000 99919991

• Note that in order to produce 16 digit integer, we must produce 2 integers of 8 size each, and concatenate them in a string. This is because of the size limit of the Java.lang.Integer class in Java.

Built-in transformation functions

Email

The Email built-in transformation function does the following to the format string output:

• Letters are lowercased: A-Z->a-z
• Numbers and lowercased letters are unchanged
• Some country specific letters are replaced:
  • æ->ae
  • ø->oe
  • å->aa
• Some symbols are replaced by a dot:
  • (whitespace) ->.
  • -->.
  • _->.
  • ..->.
• All other characters are removed
• Checks if string contains @:
  • if not, then append @mail.com to the end of the string.
  • if last character, then append mail.com to the end of the string.
  • else, nothing is appended.

CreditCard

The CreditCard built-in transformation function uses the Luhn algorithm for Credit Card encoding. It replaces the last digit for MOD10 checksum validation.

• The format string must be digits for the transform to work, e.g., "1234567890123456".

Custom Transformations

If you want to write your own transformation, see Custom Transformations for details.