通过正则表达式查找字符串中的十进制数

时间:2011-08-15 21:29:50

标签: php regex decimal

我解析命令时遇到小数问题。 12.90 已被错误地解析:

// My command
$cmd = 'CREATE product price:12.90, name: "Create me"';

// My parser
preg_match_all('/\w+|".*?"|(?!\s)\W/', $cmd, $list);

输出:

Array
    (
        [0] => CREATE
        [1] => product
        [2] => price
        [3] => :
        [4] => 12 // << problem starts here
        [5] => .
        [6] => 90
        [7] => ,
        [8] => name
        [9] => :
        [10] => "Create me"
    )

我正在寻找这个输出:

Array
    (
        [0] => CREATE
        [1] => product
        [2] => price
        [3] => :
        [4] => 12.90 // supposed
        [5] => ,
        [6] => name
        [7] => :
        [8] => "Create me"
    )

那么,我该如何解决这个问题?

编辑:(更好的解决方案)

伙计们,我根据我的问题接受了杰西的回答。但我意识到这对我的复杂命令来说还不够好。它更好,因为它适用于12.90这样的小数p.price这样的别名。所以请看下面的示例和解析器吧。我希望这有助于某人。

// My command
$cmd = 'GET `order` -o, product -p 
           LIST o.user_id, o.product_id, p.name, p.price
           REL o.order_id = 3 AND p.price > 12.90'; 

// My complex command:
preg_match_all('/[0-9_\.]+|\w+|".*?"|`.*?`|\'.*?\'|!=|<=|>=|(?!\s)\W/', $cmd, $list); 

// Output:  
Array
(
[0] => Array
    (
        [0] => GET
        [1] => `order`
        [2] => -
        [3] => o
        [4] => ,
        [5] => product
        [6] => -
        [7] => p
        [8] => LIST
        [9] => o
        [10] => .
        [11] => user_id
        [12] => ,
        [13] => o
        [14] => .
        [15] => product_id
        [16] => ,
        [17] => p
        [18] => .
        [19] => name
        [20] => ,
        [21] => p
        [22] => .
        [23] => price
        [24] => REL
        [25] => o
        [26] => .
        [27] => order_id
        [28] => =
        [29] => 3
        [30] => AND
        [31] => p
        [32] => .
        [33] => price
        [34] => >
        [35] => 12.90
    )

)

2 个答案:

答案 0 :(得分:5)

\w类不包含.,因此“快速”修复是按以下方式添加:

  preg_match_all('/[\w.]+|".*?"|(?!\s)\W/', $cmd, $list);

对于允许的时间段,您有其他要求吗? (即仅在数字之间匹配?)。如果是这样,您可能需要使匹配更专业。

对我来说最重要的事情是,如果句点将用作解析器的其他方面的分隔符。如果您提供的RE是您的整个解析器,那么您可能确定。

答案 1 :(得分:2)

这对我有用:

/[a-zA-Z0-9_\.]+|:|,|".*?"/