PDO vs. MySQLi

[thr]

@World_Wide_Weird

It really comes down to an aesthetic preference and how much you enjoy typing

You’ll have to excuse me but what does my enjoyment of typing having to do with this? Or are you suggesting that using a procedural database interface requires more typing then a OO-one?

@World_Wide_Weird

I prefer object-oriented code, and I always hated having to write “or die(mysql_error())” at the end of every line of code. It’s just a hunch, but I’d guess that catching all of the errors in a single catch block has to be more efficient, mainly because there’s less code to parse.

The parsing speed is such an abysmal part of total execution time I hardly see that as a variable to take into account when selecting my database interface, also I have a rather good hunch that OO-code is a fair bit slower then procedural code in php (even though it did get a performance boost with php5 afaik) considering that much of the object model was tacked on later (php5, etc.).

@Mastodont

What about stored procedures and views with parameters?

Yeah, I was wondering this to - I know MySQLi supports this, how about PDO? Anyone got a clue?

@World_Wide_Weird

The thread’s author mentioned benchmarks in his initial post, from which one might reasonably infer that speed and performance are among his concerns. That is the matter I addressed. Supposing that an extension that lends itself well to a procedural style of coding (and yes, I do realize that mysqli provides for OOP-style usage, although, in my opinion, it leaves much to be desired) is faster and more efficient than an extension designed for OOP coding, it is my contention that the overhead of parsing the extraneous (and, I believe, unnecessary) procedural-style error handling code might offset or even completely negate such benefits. That is all I was saying, really.

As I stated earlier, OO code is generally slower then the same procedural code in php, OO code also tends to (in the end, not in short examples) be bigger (in amount of characters), but as I said - the parsing speed is such an abysmal part of the entire execution in time that it doesn’t really matter.

@World_Wide_Weird

But since you brought up the need to maintain one’s application in a structured manner, am I to infer that you think handling individual errors as one goes along exhibits better structure than trying a functional unit of code and then catching any exceptions if the unit fails? Or would you say that one approach is neither better nor worse than the other, and that it is solely a matter of preference?

I don’t get this obsession with try/catch, since I will wrap either MySQLi or PDO in an object of my own, all errors reported from MySQLi will of course be thrown as exceptions of my choosing as all PDO exceptions will be caught within my wrapping class, re-formated and re-thrown.