Question

我有一个文件（注意：它不是一个json文件）我需要解析并插入到数据库中，因此大括号内以#TRANS开头的行应该属于以#VER开头的行以上。我想我应该使用preg_match或preg_match_all但不确定正则表达式应该如何？

#VER 1 5 20170128
{
   #TRANS 8000 {} 4016.00 20170128 "something"
   #TRANS 1100 {} -4016.00 20170128 "something"
}
#VER 1 6 20170128
{
   #TRANS 8010 {} 5016.00 20170128 "something else"
   #TRANS 1130 {} -5016.00 20170128 "something else"
}

我会将它们分成两个表，我可以分多步完成，而不是一次性完成。所以首先我要插入VER行，然后在花括号内的下方获取TRANS行，并遍历找到的行并解析它们。

以下是从文件的第一个示例开始，DB的示例：

Table VER:
ID | VER_DATE | (some other stuff)
5  | 20170128 | ...

Table TRANS:
ID | VER_ID | SERIAL | VALUE |   DATE
1  |    5   |  8000  |  4016 | 20170128
1  |    5   |  1100  | -4016 | 20170128

Answer 1

好吧，如果您不需要支持复杂的语法并且只进行一次转换，那么只需使用fopen和fgets即可轻松完成：

<?php

$fp = fopen('/path/to/file', 'rb');

$head        = '';
$blockBuffer = '';

while (true) {
    $line = fgets($fp);

    if ($line === false) break;

    if (substr($line, 0, 4) === '#VER') {
        $head        = $line;
        $blockBuffer = '';
    }

    if (!empty($head)) {
        $blockBuffer .= $line;
    }

    if (trim($line) === '}') {
        var_dump($head);
        var_dump($blockBuffer);

        $head = '';
    }
}

请注意，这是解析数据的脆弱方式。

Answer 2

您可以通过解析和处理每一行来采用某种编译器方法。它具有处理格式错误的数据的能力。

class ParserException extends Exception {
    public $line;
    function __construct($message, $line) {
        parent::__construct($message);
        $this->line = $line;
    }
}

class Parser {

    const Top = 1;
    const InBlock = 2;
    const Ver = 4;
    const Noop = 6;

    public $ver = null;

    function parse($input) {
        $lines = explode("\n", $input);
        $state = self::Top;
        $this->line = 1;
        try {
            foreach($lines as $line) {

                // echo "$line\n";
                switch($state) {
                    case self::Top:
                        if ( preg_match('/^#VER (\d) (\d) ([0-9]+)/', $line, $matches) ) {
                            $this->emitVer($matches);
                            $state = self::Ver;
                        }
                        else {
                            throw new ParserException("VER not found", $this->line);
                        }
                        break;
                    case self::Ver:
                        if ( substr(trim($line), 0, 1) == '{' ) {
                            $state = self::InBlock;
                        }
                        else {
                            throw new ParserException("Expected { ", $this->line);
                        }
                        break;
                    case self::InBlock:
                        $trimline = trim($line);
                        // echo ">>> $trimline\n";
                        // #TRANS 8000 {} 4016.00 20170128 "something"
                        if ( preg_match('/^#TRANS ([0-9]+) \{\} ([0-9.-]+) ([0-9]+) "(.*)"/', $trimline, $matches) ) {
                            $this->emitTrans($matches);
                        } elseif ( substr($trimline, 0, 1) == '}' ) {
                            $state = self::Top;
                        }
                        else {
                            throw new ParserException("Expected TRANS or } ", $this->line);
                        }                        
                        break;
                    default:
                        // unknown state
                        throw new ParserException("Unexpected error ", $this->line);                        
                        break;
                }
                $this->line++;
            }
        }
        catch(ParserException $e) {
            echo "Parser error. " . $e->getMessage() . ' Line ' . $e->line . PHP_EOL; 
        }
    }

    function emitVer($ver) {
        echo sprintf("id %s version date %s\n", $ver[2], $ver[3]);
        $this->ver = ['id' => $ver[2], 'date' => $ver[3]]; // remember the version 
    }
    function emitTrans($trans) {
        echo sprintf("Trans ver-id = %s serial = %s value = %s date = %s\n", $this->ver['id'], $trans[1], $trans[2], $trans[3], $trans[4]);
    }

    function outr($x) {
        print_r($x);
        echo "\n";
    }
}

$p = new Parser;
$p->parse('#VER 1 5 20170128
{
#TRANS 8000 {} 4016.00 20170128 "something"
#TRANS 1100 {} -4016.00 20170128 "something"
}
#VER 1 6 20170128
{
#TRANS 8010 {} 5016.00 20170128 "something else"
#TRANS 1130 {} -5016.00 20170128 "something else"
}');

Answer 3

如果您想使用正则表达式，可以这样做：

preg_match_all('/#VER (.*)\s*\{((\s*#TRANS\s*[0-9]+\s\{.*\}.*$)*\s*)\}/msU', $text, $matches);

$size = sizeof($matches[0]);

for($i = 0; $i < $size; $i++){
    $text = $matches[2][$i];
    echo "================\n";
    echo $matches[1][$i]."\n";
    preg_match_all('/\s*#TRANS\s([0-9]+)\s(\{.*\})\s([0-9\-\.]+)\s([0-9]+)\s"(.*)"$/', $text, $matches2);
    $size2 = sizeof($matches2[0]); 
    for($j = 0; $j < $size2; $j++){
        echo $matches2[1][$j]."\n";
        echo $matches2[2][$j]."\n";
        echo $matches2[3][$j]."\n";
        echo $matches2[4][$j]."\n";
        echo $matches2[5][$j]."\n";
    }
}

输出将是：

================
1 5 20170128
1100
{}
-4016.00
20170128
something
================
1 6 20170128
1130
{}
-5016.00
20170128
something else

如果您愿意，可以拆分标题($matches[1][$i])。

在PHP中用花括号括起来

3 个答案: