Data Masking

[BSG1]

Hi All,

I need to mask varchar data. How can we achieve it without using data masking stage in datastage parallel job?

I need some process through which I can unmask it also later on if needed.

Can we use convert function on individual letters?

[chulett]

Why don't we start by you defining what you mean by "mask"? Examples are always good. I've found there's a lot of confusion out there about the term and sometimes people really mean encrypt or redact. Depends on how you'll be using the data afterwards. If you need to reverse the process, you're probably not talking about masking as it is not possible to undo it.

[BSG1]

for example if my column has value A123456

it should be masked and presented as may be X456789

no particular sequence. It's just that original actual data shouldn't be present.

[chulett]

What about names? Phone numbers? Addresses? Will you be dealing with data like that? I'm also concerned this whole topic might be a non-starter if you think you need to be able to reverse the process.

[BSG1]

no

only one column which has alphanumeric data

[chulett]

If that's literally all you need to do then perhaps a simple cypher like ROT13 would work for you? Or some other letter substitution cypher of your own design? Both of those can be reversed. Keep in mind though, the simpler it is for you to build the simpler it would be for someone to break. If the security of this process / data is a major concern, you'll need to do something more complex.

[BSG1]

Can you help how to implement ROT13 in data stage?

[chulett]

It's just a lookup table which lends itself to the Convert() function rather nicely. Or you could code something in C++ or even BASIC and use a Server job. All kinds of ways.

And there's absolutely no need to quote everything each time you reply, there's a perfectly lovely Reply to topic 'button' you should be using instead.

[BSG1]

Thanks. would look into it.