Unless it is required to make heavy calculations to fill object's content, creation is faster. Note that retrieving from database also includes object creation, and access to database itself requires a lot of processor's cycles.
Alexei Kaigorodov wrote:Note that retrieving from database also includes object creation, and access to database itself requires a lot of processor's cycles.
Although I agree with you in general, it's worth noting that the whole reason that databases exist is to persist and retrieve data; so, while for an individual object what you say is right, if Paras needs to store lots of objects, a database (NOT a file) is probably the way to go.
@Paras: Either way, serialization, while convenient, is extremely slow. It's more often used with Streams to pass entire objects across a network than as a primary persistence mechanism. And, unless you're dealing with very small amounts of data, a database is probably better than using files - in fact, with a 'wrapped' db like JavaDB it may even be better for individual objects too.
Isn't it funny how there's always time and money enough to do it WRONG?
Joined: May 22, 2012
Well, i only know how to store objects in a file. Let me know how to do the same in database.
Also, another performance related question:
Is it better to import whole packages or just the classes we need to use.
I am making quite a big project and such minute things are affecting my performance (i think)
fred rosenberger wrote:
So having a lot of imports may slow down your compile phase, it will have ZERO impact on your run-time performance.
(at least, this is what I have read many, many times).
You can prepare a test case: two versions of a class, one with fully qualified class names, the other the same but with imports,
and then you can disassemble the resulting classes to verify they are essentially identical.