从查询正则表达式中提取值

Question

I need to extract the values ​​of a condition (WHERE) and did a regex, but I can not get the values ​​correctly.

//Patherns
$regex  = "/([a-zA-Z_]+)\s([\<\=\>\s]{0,4})\s+(\".*\")/";

//values ​​to be extracted
$string = 'idCidade >= "bla" OR idEstado="2" and idPais="3"'; 

//regex function
preg_match_all(
    $regex,
    $string,
    $output
);

//displays the result
echo '<pre>';print_r($output);



//incorrect output
Array
(
   [0] => Array
    (
        [0] => idCidade >= "bla" OR idEstado="2" and idPais="3"
    )

   [1] => Array
    (
        [0] => idCidade 
    )

   [2] => Array
    (
        [0] => >= 
    )

   [3] => Array
    (
        [0] => "bla" OR idEstado="2" and idPais="3"
    )
)

I need the regular expression to export the values ​​to an array like this;

//correct output
Array
(
   [0] => Array
    (
        [0] => idCidade >= "bla" OR idEstado="2" and idPais="3"
    )

   [1] => Array
    (
        [0] => idCidade
        [1] => idEstado
        [2] => idPais
    )

   [2] => Array
    (
        [0] => >=
        [1] => =
        [2] => =
    )

   [3] => Array
    (
        [0] => "bla"
        [1] => "2"
        [2] => "3"
    )
   [4] => Array
    (
        [0] => "OR"
        [1] => "AND"
        [2] => ""
    )
)

Answer 1

Your mistake was probably the .* which matches too much. You'd need to make it "ungreedy" with appending a question mark: .*?

I would however suggest this regex:

'/(OR|AND)?\s*(\w+)\s*([<=>!]+)\s*("[^"]*"|\'[^\']*\'|\d+)/i'

This matches the boolean connector first and optionally, so that you get:

[1] => Array
    (
        [0] => 
        [1] => OR
        [2] => and
    )

[2] => Array
    (
        [0] => idCidade
        [1] => idEstado
        [2] => idPais
    )

[3] => Array
    (
        [0] => >=
        [1] => =
        [2] => =
    )

[4] => Array
    (
        [0] => "bla"
        [1] => "2"
        [2] => "3"
    )

I've also made it work for SQL-compliant strings and decimals. But this is only borderline a job for regex. A real parser would be advisable. (Though I don't know your use case.)

Answer 2

Try this. This outputs the exact result you need.

<?php  //Patherns
$regex  = '/([a-zA-Z_]+)\s*([>=<]*)\s*"([^"]*)"\s*(or|and)*/i';

//values to be extracted
$string = 'idCidade >= "bla" OR idEstado="2" and idPais="3"';

//regex function
preg_match_all(
    $regex,
    $string,
    $output
);

//displays the result
echo '<pre>';print_r($output);