Preg_split question

[pascalito]

Hi,

i have a problem with a regular expression.

i want to split a string on character pipe | with preg_split php function.

The string is :

"foo|bar|'hello |world\'abc\'de f'|bye"

Result of split must be in an array :

foo
bar
hello|world'abc 'def
bye


Anyone have an idea ?
Thank you !!
bye

[Markus]

What is the expression you're using?

[pascalito]

What is the expression you're using?

i use preg_split function like this :

$val = "foo|bar|'hello|world\'abc\'def'|bye";
$val = preg_split('/(?:\s\|\s)(?![\w\s]*\')/', $val, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY);
print_r($val);

But the pattern is wrong...

[Markus]

Hmm, the delimiter of your string is simply '|', that is, a pipe character? If so, explode() would be much better suited to it.

$string = "mark|jon|james|sarah|rachel|ben";

print_r(explode('|', $string));

[pascalito]

Hmm, the delimiter of your string is simply '|', that is, a pipe character? If so, explode() would be much better suited to it.

$string = "mark|jon|james|sarah|rachel|ben";

print_r(explode('|', $string));

No, watch my example :

"foo|bar|'hello |world\'abc\'de f'|bye"

With explode function, i obtains :

foo
bar
'hello
world\'abc\'def '
|bye

I would like :

foo
bar
hello|world'abc 'def
bye

It's not the same...
Thank you

[TheServant]

I can't imagine what your string is actually about. Why is hello|world'abc 'def surrounded by inverted commas?? You would also like a pipe character | in one of the lines (not deliminated)?
I have a feeling that you will want to use regex functions rather than preg_split or explode(), and explicitly state the exact pattern...

It would be much easier to change your initial input to make the real deliminator something not included in one of your lines, or even a pattern like ||:

print_r(explode("||","foo||bar||'hello|world\'abc\'def'||bye"));

Oh, and for the \'s in the exploded version, you might try stripslashes on it to get rid of them.

[pascalito]

I can't imagine what your string is actually about. Why is hello|world'abc 'def surrounded by inverted commas?? You would also like a pipe character | in one of the lines (not deliminated)?
I have a feeling that you will want to use regex functions rather than preg_split or explode(), and explicitly state the exact pattern...

It would be much easier to change your initial input to make the real deliminator something not included in one of your lines, or even a pattern like ||:

print_r(explode("||","foo||bar||'hello|world\'abc\'def'||bye"));

Oh, and for the \'s in the exploded version, you might try stripslashes on it to get rid of them.

Your example is not good :

If my string is "foo|bar|'hello ||world\'abc\'d ef'|bye" (note the ||)

your example code return :

Array
(
    [0] => foo
    [1] => bar
    [2] => 'hello
    [3] => world\'abc\'def'
    [4] => bye
)

« 'hello||world\' abc\'def' » is splitted in 2 parts... not good

I have found a solution with preg :

$val = preg_split("/([a-zA-Z0-9]+:'.+?'|[^\|]+)\||$/", $val, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY);

Bye

[TheServant]

Glad you got it solved, but you did not read what I wrote. Between hello|world there is only one | while where you want a new line you use two ||.

You said: "foo|bar|'hello ||world\'abc\'d ef'|bye"
I said: "foo||bar||'hel lo|world\'abc\' def'||bye"

Do you see the difference? I still think that as Markus suggested, it would be quicker and easier to use explode(), but as long as it works, I guess you're happy.

[pascalito]

Glad you got it solved, but you did not read what I wrote. Between hello|world there is only one | while where you want a new line you use two ||.

You said: "foo|bar|'hello ||world\'abc\'d ef'|bye"
I said: "foo||bar||'hel lo|world\'abc\' def'||bye"

Do you see the difference? I still think that as Markus suggested, it would be quicker and easier to use explode(), but as long as it works, I guess you're happy.

quicker and easier ??
ok, write me the function who parse input data...

[pascalito]

there is a mistake in the preg_split solution, pattern is "/('.+?'|[^\|]+)\||$/"

[TheServant]

quicker and easier ??
ok, write me the function who parse input data...

Are you controlling your deliminators or are they changing??
I wrote a code:

print_r(explode("||","foo||bar||'hello|world\'abc\'def'||bye"));

/* [U]Outputs[/U]: */
Array 
( 
    [0] => foo 
    [1] => bar 
    [2] => 'hello|world\'abc\'def' 
    [3] => bye 
)

If you throw in a stripslashes:

print_r(explode("||",stripslashes("foo||bar||'hello|world\'abc\'def'||bye")));

/* [U]Outputs[/U]: */
Array 
( 
    [0] => foo 
    [1] => bar 
    [2] => 'hello|world'abc'def' 
    [3] => bye 
)

You have not shown us how your data is fetched/made, all you said was how do you turn "foo|bar|'hello |world\'abc\'de f'|bye" into and array of:

foo
bar
hello|world'abc 'def
bye

I suggested that if you change your deliminator from | to || then you can use explode while keeping your hello|world together.
If you deliminator is changing, you did not mention that. If you find preg_split easier, go for your life. Generally explode and similar string functions are easier for debugging.

[pascalito]

TheServant, you do not answer the question...

I know how to use the explode function, everyone knows how to use this function...

you say it is quicker and easier to convert input data, ok, I ask you to write the function that converts input data from "foo|bar|'hello |world\'abc\'de f'|bye" to "foo||bar||'hel lo|world\'abc\' def'||bye"

Regards,

[pascalito]

I do not control the delimitator, otherwise, I never asked and I simplified input string...

[TheServant]

You never said that you couldn't control the deliminator, and how am I meant to know your experience? In that case, using regex (or preg_split) is the way you should go. Glad it's solved.

[pascalito]

You never said that you couldn't control the deliminator, and how am I meant to know your experience? In that case, using regex (or preg_split) is the way you should go. Glad it's solved.

I never said any more than I could change the string but it seems logical to me.

If I could change the string, do you really think I asked any questions ...