[RESOLVED] Remove begining of a string.

Sops21

I have what should be something really simple to solve. I have the following string:

tag:search.twitter.com,2005:950308934

The only information that I need from this string is

950308934

Basically I need to remove everything before and including the ":" . I'm pretty sure that this should be a simple problem to solve. I just can't seem to find the php function to help me with this.

Any suggestions?

NogDog

If the part at the end you want is always the same number of digits, you could use [man]substr/man with a negative number for the offset parameter.

Otherwise you could do an [man]explode/man on the ":" and the [man]array_pop/man to grab the last element of the resulting array.

nrg_alpha

If you know the last chunk will only be numbers, you could do this:

$str = 'tag:search.twitter.com,2005:950308934'; 
preg_match('#(\d+)$#' , $str, $match);
echo $match[1];

Output:

950308934

Weedpacket

If the number of digits can vary, you can use [man]strrpos[/man] to find the last ':' and then use [man]substr[/man] to capture everything after that point.

nrg_alpha

Or you can use the preg pattern supplied 😉

Weedpacket

Yah, but it's generally more efficient to use one or two of the more conventional functions rather than breaking out regex at the first opportunity. Incidentally, did you know that ¹, ², and ³ count as digits to PCRE?

nrg_alpha

Weedpacket;10888972 wrote:
Yah, but it's generally more efficient to use one or two of the more conventional functions rather than breaking out regex at the first opportunity.

Perhaps, but for small tasks like this, its like splitting hairs as the 'performance differences' between those methods are negligible at worst. If the task was to check through much larger amounts of data (or loop through this kind of calculations heavily), then I agree.

Weedpacket;10888972 wrote:
Incidentally, did you know that ¹, ², and ³ count as digits to PCRE?

I did not know that ( something new everyday they say 😉 ). I would lump this kind of thing in with \w being capable of matching é for example (depending on locale). It's not something many people should concern themselves much with. If need be, in this case, you could simply use ([0-9]+) instead.

Sops21

Thanks for all your help.
I chose to use this solution.

$str = 'tag:search.twitter.com,2005:950308934';
preg_match('#(\d+)$#' , $str, $match);
echo $match[1];

Weedpacket

nrg_alpha wrote:
but for small tasks like this

It's precisely small tasks like this for which regular expressions are overkill. By your argument one would never use strpos, substr, trim, or strspn because either it's either such a small task that it wouldn't make any difference (and the regular expression is probably more complicated than necessary) or such a large task that it would be more legibly written as a regular expression.

E.g., if it's always nine digits:

$match = substr($str,-9);

As far as speed goes (which isn't my only criterion for efficiency),

substr($str,strrpos($str,':')+1)

is faster than that regular expression by more than half an order of magnitude, which is my personal threshold for thinking that it matters. Incidentally, that regular expression would be somewhat faster without the () in it.

nrg_alpha

Weedpacket;10889075 wrote:
It's precisely small tasks like this for which regular expressions are overkill. By your argument one would never use strpos, substr, trim, or strspn because either it's either such a small task that it wouldn't make any difference (and the regular expression is probably more complicated than necessary) or such a large task that it would be more legibly written as a regular expression.

In cases like this, the lag performance would be negligible. In this particular case, for a simple search through a simple string, preg wont't break the performance bank if you will.

I'm not suggesting no one use substr, str_replace or the like. If they prefer to go that route, go for it.. This is the method I would choose, overkill or not.

Sure, preg patterns run the risk of not being as well written as well as they can be a) I have more than my share of that, and
b) certainly will hurt the speed performance

But as you learn more and become more experienced, you start writing better patterns, and the differences starts to diminish to an extent (when I use the term you, I am not literally referring to you, weedpacket.. just you as in people in general).

Right or wrong, this is what I chose and offered. If the OP sees serious lag in performance as a result, then I urge going another route. It's that simple.