The key win of relational databases is that they allow people to store, manage, manipulate and flexibly query data structures without having to think in detail about algorithms. If I was managing a few hundred records and needed to do things like easily find what classes Johnny took, I would be inclined to use a relational database for those features. And if I had a good one, it would grow with me if I needed to handle millions of records without my having to get into some heavy-duty wizardry.
However the problem of efficient storage and access to data is independent of the data management structure of a relational database. That is the role of a dbm. The technologies in dbms are buried somewhere inside of traditional relational databases. (MySQL can actually run on top of Berkeley DB.) But sometimes your structure is simple enough that you can manage it yourself, and in that case there is no reason not to remove the relational layer. Large amounts of data is not the same as complex data.
(There are also other management strategies than relational, for instance object oriented databases like Gemstone.)
In reply to Re (tilly) 2: millions of records in a Hash
by tilly
in thread millions of records in a Hash
by johnkj
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |