涵盖多行的正则表达式

Question

I would like to extract the data from the following sample txt

START
A: 567 B:345
C: 345 D:485


START
A: 345 B:267
C: 980 D:054

I'm using this regex with no positive results:

$regex = '~'
       . '^(?P<header>.*START:.*)$'
       //. '^(.*A: \d{3} B: \d{3}.*)$'
       //. '(.*C: \d{3} D: \d{3}.*)$'
       . '~m'
;

That gives me the following result:

Array
(
    [0] => Array
        (
            [0] =>  START:
            [header] =>     START:
            [1] =>  START:
        )

    [1] => Array
        (
            [0] =>  START:
            [header] =>     START:
            [1] =>  START:
        )

)

But as soon as I uncomment the lines

. '^(.*A: \\d{3} B: \\d{3}.*)$'

and

. '(.*C: \\d{3} D: \\d{3}.*)$'

I get no match!

How can I extract this data?

I've tried deleting the start of line anchor ^ and changing the end of line anchor $ to \\n with no success

IN GENERAL

How can I extract the data when I have a start of data line followed by n data lines, followed by a end of data line followed by one or more empty lines?

Answer 1

Try this:

(START)\s+(A:\s+\d+\s+B:\d+)\s+(C:\s+\d+\s+D:\d+)

---

MATCHES:

MATCH 1
1.  [0-5]   `START`
2.  [6-18]  `A: 567 B:345`
3.  [19-31] `C: 345 D:485`
MATCH 2
1.  [34-39] `START`
2.  [40-52] `A: 345 B:267`
3.  [53-65] `C: 980 D:054`

---

Regex101 Demo
Ideone Demo

Answer 2

While matching multiline string with modifier m (MULTILINE , you have to make sure you are matching following newline with [\\r\\n] or better with \\R .

For your input following regex should work:

/^(?<header>START)$\R+^(?<line1>A:\h*\d{3}\h+B:\h*\d{3})$\R+^(?<line2>C:\h*\d{3}\h+D:\h*\d{3})$/m

\\R matches any newline and \\h matches any horizontal whitespace in PCRE.

RegEx Demo