Re: Big O

Andreas is correct but there is more to it than that. I am assuming that
you are writing a paper for a computer science course and this would not be complete enough.

When you mention search times what type of searches are you referring to. In the case of a single vaiable search, to find the where the physical data is located is order (n log n). Once you go to the disk then it is a matter of collecting the data. Is the data ordered on the disk (clustered ) or is it fragmented and going to cause disk clobbering. (seeks). The difference here is that the DBMS or the O/S has the control over putting the files on the disk. The DBMS can ask for them to be clustered but the O/S will try but if there is not a continguous space for the files then the become scattered.

So the retrieval is another question. There are many advantages of a DBMS over a flat file system and I am going to name a few and if you want to post more speifics then I can anwer them in more detail.

1)Data independence - apps work independent from the data in a DBMS as all the control of the data is in the hand of the DBMS where in the FFS(flat file system) you can perfrm your own buffering.

2) data access - More efficient in DBMS becuase technologies are specifically deployed for that reason.

3) data integrity and security - use of constraints in the database for data, again away from the program

4) admin - centralization of data so that it can be easily administered ... minimal redundancy and easy fine tuning and easy for changes.

5) concurrent access and crash recovery - people can think of accessing data as if they are the only one. protect these users fom crashes

6) reduce app development time - apps can focus on other things and can become more robust as data is being handled by the DBMS and a simple enough interface is provided.

a DBMS is not better when an app needs to be very very efficient where incorporating some features of a DBMS in the program is more desirable. Rememeber to find the data could be n log n but the creation of the query can take a long time. app needs to manipulate data that is not available in the DBMS.

Let me know if you want it to be more scientific and I can help. I use to teach a course on this.

Eric

Nashirak Bosk <bluhmcc_at_auburn.edu> wrote in message news:<u54i63m3gt0628_at_corp.supernews.com>...
> I am writing something on the advantages of Databases over flatfile
> systems in most Commercial Situations. Does anyone know the search times
> (in the order of Big O)?? I am thinking its pretty close to O(1)
> (Through hashing) but does anyone know about the average big O for most
> Databases?
Received on Sun Jan 27 2002 - 04:11:14 CET