Question

I have a PHP application that is supposed to parse uploaded text files that has a format similar to this:

|                  |                |                  |
| -----------------------------------------------------|
| Sample           | Data           |                  |
| -----------------------------------------------------|
| Sample           | Data           |                  |
| -----------------------------------------------------|
| Sample           | Data           |                  |
| -----------------------------------------------------|


| Accepts                    |                            |
| --------------------------------------------------------|
| All                        | Yes                        |
| --------------------------------------------------------|
| More                       | Yes                        |
| --------------------------------------------------------|


|            |            | Years      |            |            |
| ---------------------------------------------------------------|
| 1998       | 1999       | 2000       | 2001       | 2002       |
| ---------------------------------------------------------------|
| 2003       | 2004       | 2005       | 2006       | 2007       |
| ---------------------------------------------------------------|
| 2008       | 2009       | 2010       | 2011       | 2012       |
| ---------------------------------------------------------------|

What I need to do is basically isolate each “block” by itself in the same order, so I can loop them one-by-one. A “solution” could be doing

preg_split("/\n{4,}/", $text);

However that would produce unwated results if the person submitting the text decides that the unnecessary newlines doesn’t belong and removes them. I tried playing around with preg_match_all(), but it has been years since I did any real regex, so I couldn’t come up with a usable solution.

The first line of a “block” always contains | and spaces, but fields may contain text. The last line of a “block” is always a pipe followed by a space, dashes to fill the row, ending with a |.

Answer 1

If this is how the content of the text file looks like I would write something like

$pat = '~
    (?<=^|\r{3}|\n{3}|(\r\n){3})    # beginning of string or following 3 newline chars
        \|[ ]                       # a pipe and a space
        (
            [ \S]+                      # 1 or more space or non space char
            \|                          # a pipe
        )+                          # 1 or more of this group

        (\n|\r\n?)                  # a newline
        \|[ ]-+\|                   # a pipe, a space, multiple dashes and a pipe
        (\n|\r\n?)                  # a newline
        .*?                         # anything between newlines above and below
        (\n|\r\n?)                  # a newline
        \|[ ]-+\|                   # a pipe, a space, multiple dashes and a pipe
    (?=$|\r{3}|\n{3}|(\r\n){3})     # end of string or followed by 3 newline chars
~sx';
preg_match_all($pat,$str,$res);
$blocks = $res[0];
print_r($blocks);

I’m not sure if this is the most elegant or even reliable way, though, since it’s hard to guess what exactly the content might look like.