If your programming language permitted the development of database tables, only permitted one area for use being an index ... how restricting would that be? I'm not a genuine database programmer, so I am wondering if this type of system could be utilized for the foundation for serious databases, or could be merely a toy.

That might be very restricting. I have seen as much as 11 indexes on tables I have worked with, these getting an objective. Without doubt you will find cases available with substantially more.

At least inside a database the main key and all sorts of foreign secrets is going to be (or ought to be) indexed.

In fact it is common for indexes to span multiple posts.

Non-relational databases (presently extremely popular using the NoSQL movement) approximate this type of situation. Many such databases are not toys, and therefore are intensely used because the foundation of greatly scalable production systems -- the possible lack of free indexing along with other restrictions empower almost unbelievable scalability, specifically for "mostly-read" situations. However essentially in needing to quit normalization, and copying information everywhere with techniques which are strictly depending on what queries need what amount of optimisation.

I am unsure what your "single index" scenario is dependant on, but unless of course it provides exactly the same types of advantages because the best "nosql" storage systems (and you are prepared to fully spend the money for cost for your), it appears unlikely to become a technically viable concept-).

It is dependent. Berkeley DB e.g. has only just one key and it is very helpful because of its reasons.

A relational DB however should provide you with the chance to define as numerous indices as you desire (as cletus pointed out, foreign key queries would are afflicted by inadequate performance otherwise)

Generally: a catalog is perfect for accelerating database research: as with a magazine, you search the term "computer", you see the (alphabetical!) index, and discover all page-amounts.

If (very hypothetical reasoning) one produces a database which enhanced research occasions without needing indices: then this type of database that contains just one index isn't a bad idea.

But when you are speaking in regards to a mainstream DB, then I'd say: is dependent around the content.

It is dependent on what you're going related to the information.

Should you search and choose according to an unindexed area, the procedure will probably be slow. Likewise, should you perform join or merge procedures according to an unindexed area, you receive a decelerate.

When the table has countless rows, this may be not fast enough to become acceptable.

Restricting you to ultimately access by means of only one area can in a few instances be too restricting.

Again, it is dependent on what you're going related to the information.