GDPR - string anonymisation function

[mLipok]

I need to anonymize some strings.

For example:
 

_Example()
Func _Example()
    Local $sString1 = 'Warsaw'
    MsgBox(0, 'Test 1a', _GDPR_Anonymize_String1($sString1))
    MsgBox(0, 'Test 1b', _GDPR_Anonymize_String2($sString1))

    Local $sString2 = 'Zielona Góra'
    MsgBox(0, 'Test 2a', _GDPR_Anonymize_String1($sString2))
    MsgBox(0, 'Test 2b', _GDPR_Anonymize_String2($sString2))

    Local $sString3 = 'ZAKŁAD UBEZPIECZEŃ SPOŁECZNYCH'
    MsgBox(0, 'Test 3a', _GDPR_Anonymize_String1($sString3))
    MsgBox(0, 'Test 3b', _GDPR_Anonymize_String2($sString3))

    Local $sString4 = 'ZAKŁAD UBEZPIECZEŃ SPOŁECZNYC'
    MsgBox(0, 'Test 4a', _GDPR_Anonymize_String1($sString4))
    MsgBox(0, 'Test 4b', _GDPR_Anonymize_String2($sString4))
EndFunc   ;==>_Example


Func _GDPR_Anonymize_String1(ByRef $sString)
    Return SetError(@error, @extended, StringRegExpReplace($sString, '(?i)(.)(.{1,2})', '$1**'))
EndFunc   ;==>_GDPR_Anonymize_String1

Func _GDPR_Anonymize_String2(ByRef $sString)
    Return SetError(@error, @extended, StringRegExpReplace($sString, '(?i)(.)(..)', '$1**'))
EndFunc   ;==>_GDPR_Anonymize_String2

 

What is your opinion about this example, and compliance with GDPR  "guidelines"?
Maybe you have a better solution to this problem?

Edited February 28, 2019 by mLipok

[abberration]

Never heard of GDPR, but I found this article that might help:

https://gdpr.report/news/2017/09/28/data-masking-anonymization-pseudonymization/

It looks like data masking is allowed, but the article didn't specify how much of it needs to be masked.

[FrancescoDiMuro]

@mLipok

Go for the data pseudonymisation.

It can be applied to large amount of data, and it is more reliable than the anonymisation, since, in the first case, data is protected/crypted, while in the second case, data is "changed", and the work to rebuild it is significantly bigger than using the first method 

[Danyfirex]

Hello. Write it in that way you will not affect @error and @extented. 

Saludos

[mLipok]

17 hours ago, abberration said:

Never heard of GDPR

 

Are you from USA?

16 hours ago, FrancescoDiMuro said:

Go for the data pseudonymisation......and the work to rebuild it is significantly bigger than using the first method

1

My intention was to create a function which can put some restricted data in a "save form" to TXT log Files or other "output".

#include "ErrorLog.au3" ; https://www.autoitscript.com/forum/index.php?showtopic=195882
_Log_SetOutputFunction(ConsoleWrite)

_Example()

Func _Example()
    Local $sString2 = 'Zielona Góra'
    _Log('City: ' & _GDPR_Anonymize_String1($sString2) & @CRLF)
EndFunc   ;==>_Example



Func _GDPR_Anonymize_String1(ByRef $sString, $iError = @error, $iExtended = @extended)
    Return SetError($iError, $iExtended, StringRegExpReplace($sString, '(?i)(.)(.{1,2})', '$1**'))
EndFunc   ;==>_GDPR_Anonymize_String1

Func _GDPR_Anonymize_String2(ByRef $sString, $iError = @error, $iExtended = @extended)
    Return SetError($iError, $iExtended, StringRegExpReplace($sString, '(?i)(.)(..)', '$1**'))
EndFunc   ;==>_GDPR_Anonymize_String2

 

[Danyfirex]

It still does not affect @error and @extended.

 

Saludos

[FrancescoDiMuro]

1 hour ago, mLipok said:

My intention was to create a function which can put some restricted data in a "save form" to TXT log Files or other "output".

So you don't have the need to "rebuild" the data in a later moment?
If so, then you can use the anonymisation (partial or total) 

[mLipok]

17 hours ago, Danyfirex said:

Write it in that way you will not affect @error and @extented. 

 

I do what you request (this was quite sensible)......

and

1 hour ago, Danyfirex said:

It still does not affect @error and @extended.

 

now you are asking about the opposite case?

 

[FrancescoDiMuro]

@mLipok
It means:
If you write it like you did (in both cases), you are not setting the @error code and the @extended code to what SRER returns

[mLipok]

4 minutes ago, FrancescoDiMuro said:

So you don't have the need to "rebuild" the data in a later moment?
If so, then you can use the anonymisation (partial or total) 

1

As EndUser (Client) still have the source database so this is rather a pseudonymization (from the point of view of EndUser), but if the EndUser wants to send only log data to DevTeam, then EndUser can say that Logged data was sent do DevTeam as an anonymized data.

At least ..... I HOPE THAT ⁉️

[Danyfirex]

I mean this:

_Example()
Func _Example()
    Local $sString1 = 'Warsaw'
    Local $sStringNew=_GDPR_Anonymize_String1($sString1)
    ConsoleWrite("@extended: " & @extended & @CRLF)
    ConsoleWrite("$sStringNew: " & $sStringNew & @CRLF)
EndFunc   ;==>_Example

 

@extended should be 2 in this case.

 

Saludos

[mLipok]

6 minutes ago, FrancescoDiMuro said:

to what SRER returns

What means SRER ?

[FrancescoDiMuro]

3 minutes ago, mLipok said:

As EndUser (Client) still have the source database so this is rather a pseudonymization

It depends on how data is stored in the Database.
The EndUser should be able to rebuild the data from a database in which data should be pseudonymised.
But, in this case, the EndUser has the Database with "decrypted" data, so, to send it to the DevTeam, you can just anonymise it and nothing else.

4 minutes ago, mLipok said:

What means SRER ?

StringRegExpReplace

[mLipok]

4 minutes ago, Danyfirex said:

@extended should be 2 in this case.

 

Do you mean that I should return @extended from StringRegExpReplace() function (number of replacements performed) ?

 

[Danyfirex]

Not really. I just thought You wanted that. But actually i that way you never affect @error or @extended what I want to say is that this expression:

Return SetError(@error, @extended, StringRegExpReplace($sString, '(?i)(.)(.{1,2})', '$1**'))

this part "StringRegExpReplace($sString, '(?i)(.)(.{1,2})', '$1**')" or wharever expression you run in that way will not affect @error, @extended.

Saludos

[mLipok]

@Danyfirex aha .. now I see. You were referring to my "nonsense" in the code:

Func _GDPR_Anonymize_String1(ByRef $sString)
    Return SetError(@error, @extended, StringRegExpReplace($sString, '(?i)(.)(.{1,2})', '$1**'))
EndFunc   ;==>_GDPR_Anonymize_String1

where, the code will always return      @error = 0   and    @exteneded = 0 

Yes, this was a quick beta version

 

[mLipok]

27 minutes ago, FrancescoDiMuro said:

But, in this case, the EndUser has the Database with "decrypted" data, so, to send it to the DevTeam, you can just anonymise it and nothing else.

2

So if program stores some anonymized data in a form of "Textual LOG files" ...

Question 1:
Can my function be considered sufficient to assume that the returned data is fully anonymous?

Question 2:
In the case when the answer to Q1 == Yes .... 

Whether such file is not subject to additional protection (for example I mean that you do not need to backup, etc.)  ?
or in other words:
Whether for this type of file, constituting a set of anonymous information, you still need to use any other procedures required by GPDR ?

 

[FrancescoDiMuro]

1 hour ago, mLipok said:

Can my function be considered sufficient to assume that the returned data is fully anonymous?

This article is pretty in-line with the object you are trying to achieve.
By the way, it depends about what kind of data you have; if you have a database where, querying it you can somehow rebuild the data, you should make a lot of data suppression instead of data generalization (so, make some columns * instead of W*****w).
Take a look at the article and let us know

EDIT: 
About Question 2, you have always to do backups of your data.
Through some suggested software, there's Iperius Backup.

Edited March 1, 2019 by FrancescoDiMuro