Should I store cryptographic hashes as a blob or as a string?

Question

I'm new to MongoDB and am wondering how to decide to store some cryptographic hashes. Should I store the binary data itself, as a blob, or should I store its string representation as a string?

My instinct is to just use a string and be done with it, but I'm not sure if this is the best approach. A blob would take up less space, for example.

• Is there any negative impact, coding-wise, of using a blob (i.e. harder to pass around or reference objects?)
• Are there any "gotchas" if the _id of a document is a blob?
• Is there any benefit of using a blob at all?

Answer 1

We've done it both ways and it doesn't really make much of a difference. But we've been going toward blobs on newer projects, which the PostgreSQL (our database of choice) command line shows in hex-coded format by default anyway.

The crypto APIs also tend to pass around the hashes as byte[] so blobs are a nice fit.

Finally, by using byte[] one can make an attempt to zero-out the array contents when done with the array.

Slightly off topic: noted "attempt" to zero-out because in Java objects may be moved around in memory without your control. We can zero-out an array and a previous copy of its contents may remain in physical memory and be available outside the process. We just do what we can and haven't looked into any JNI/sun.misc.Unsafe ways to pin objects.