So, if in that case it is sorted by the last name, isn't it more logical to do "Torren, Piers Titus Van der"? It seems logical to me and seems to be what happens in the case you describe, i should look at a PT list to see how they solve similar cases too
...no, that solution is the most dirty one. It's easy to implement, but it's not logical and more important: it's simply wrong.
There is a clean solution, used where personal data has to be processed professionally. That is, you have to maintain more fields in the database: additional fields for middle name, last name prefix, second name, royal title, academic title, etc.
But I think this would be overkill in our scenario.
There is a quite easy solution that gives correct results, which you can also see in music databases like in Winamp. Let me give you an example, an excerpt from my list of artists, sorted by Winamp:
- Inner City
- The Inner Space
The String, built like <last name, first name> is left as it is (so "van der Torren, Piers Titus" remains himself), but the sorting algorithm has to know a list of last name prefixes and has to ignore them when sorting. That is quite easy to achieve, gives correct results and doesn't lead to problems when working with the data.
That problem is not restricted to dutch names, but affects names from many cultures. Think of Peter O'Toole, David McAllister, etc.
A list of such last name prefixes could look like this (maybe not complete):
- aus dem
- aus den
- aus der
- de los
- van de
- van den
- van der
- vom und zum
- von dem
- von der
- von und zu