如何在转换为 json 之前验证来自 csv 的数据？

Question

我有一个这样的 csv 文件：-

+------+------+------------------------+
| name | mark |        url             |
+------+------+------------------------+
| ABCD |    5 | http://www.example.org |
| BCD  |   -2 | http://www.example.com |
| CD   |    4 | htt://www.c.com        |
+------+------+------------------------+

它包含名称、标记和网址的标题。我正在使用 PHP 将数据从 csv 转换为 json。我想在将其转换为 json 之前添加验证，就像它应该是 UTF-8 的名称一样，标记应该是一个正数，介于 0 到 5 之间，并且 url 应该是有效的。如果该行通过了所有验证，那么它将存储在 errorless.json 中，如果任何行有任何问题，则在 error.json 中并附上错误的注释。 csv转json的PHP代码：-

$fh = fopen("names.csv", "r");

$csvdata = array();

while (($row = fgetcsv($fh, 0, ",")) !== FALSE) {
    $csvdata[] = $row;
}

$fp = fopen('data.json', 'w');
fwrite($fp, json_encode($csvdata));
fclose($fp);

我想知道如何添加这些验证来转换数据。我对这些概念不熟悉，无法想出办法来做到这一点。如果有人可以帮助我，我将不胜感激。

Answer 1

这里是你要做的一个外壳：

// to make life easier on yourself create 
// a function that checks one row of data 
// to make sure it is valid
// if a row is found to be invalid, the
// function will add an `error` field to the 
// array explaining the validation error   
function validateRow(&$data) {
    if (!array_key_exists('mark', $data) && ((int)$data['mark']) >= 0 && ((int)$data['mark']) <= 5) {
        $data['error'] = "No 'mark' value found between 0 and 5";
        return false;
    }
    // do validation for $data['name']
    // do validation for $data['url']
    return true;
}

$fh = fopen("names.csv", "r");

$csvdata = array();

while (($row = fgetcsv($fh, 0, ",")) !== FALSE) {
    $csvdata[] = $row;
}
$header = $csvdata[0];
$n = count($header);
$errorless = array();
$haserrors = array();
// here is where we convert the CSV data into
// an associative array / map of key value
// pairs by treating each row and the header 
// row as parallel arrays.  Start with index 1
// in the for loop to skip over the header row
// in csvdata
for ($row = 1; $row < count($csvdata); ++$row) {
    $data = array();
    for ($x = 0; $x < $n; ++$x) {
        $data[$header[$x]] = $csvdata[$row][$x];
    }
    // if we encounter no errors, 
    // add data to the errorless array
    if (validateRow($data)) {
        $errorless[] = $data;
    }
    else {
        $haserrors[] = $data;
    }
}

// you will want to do a parallel write
// for the error.json file
$fp = fopen('errorless.json', 'w');
fwrite($fp, json_encode($errorless));
fclose($fp);

Answer 2

您应该使用一种称为 JSON 模式的东西。您定义文档的形成方式。然后可以自动验证您的文档。

如果您正在寻找另一种 CSV 验证库的具体实现，您应该首先查看 packagist.org。

Answer 3

这可能对你有帮助，

标准 JSON 格式不明确支持文件 cmets。 RFC 4627 application/json，这是一种用于存储和传输数据的轻量级格式。如果评论真的很重要，您可以将其包含为另一个数据字段，如 cmets

输入

$ cat test.csv 
name,mark,url
ABCD,5,http://www.example.org
BCD,-2,http://www.example.com
CD,4,htt://www.c.com

脚本

<?php
function validate_url($url)
{
    return in_array(parse_url($url, PHP_URL_SCHEME),array('http','https')) && filter_var($url, FILTER_VALIDATE_URL);
}
function validate_mark($val)
{
    return ($val >= 0 && $val <= 5);
}

function errors($field)
{
    $errors = array(
        'url' => 'URL should be valid',
        'mark' => 'Mark should be between 0 to 5'
    );

    return ( isset($errors[$field]) ? $errors[$field] : "Unknown");
}

function csv2array_with_validation($filename, $delimiter = ",")
{
    $header = $row = $c_row = $output = $val_func =  array();
    if (($handle = fopen($filename, 'r')) !== FALSE) 
    {
        while (($row = fgetcsv($handle, 0, $delimiter)) !== FALSE) 
        {
            if (empty($header))
            {
                $header = array_map('strtolower', $row);
                foreach ($header as $e) 
                {
                    $val_func[$e] = function_exists('validate_' . $e);
                }
                continue;
            }
            $c_row = array_combine($header, $row);
            $index = 'error_less'; $errors = array();
            foreach ($c_row as $e => $v) 
            {
                if ($val_func[$e]) 
                {
                    if (!call_user_func('validate_' . $e, $v)) 
                    {
                        $index = 'error';
                        $errors[$e] =  errors($e);
                    }
                }
            }
        /* 
          If the comment is truly important, 
          you can include it as another data field like errors,  
          comment below part if you do not wish to create new field (errors) in 
          json file 
        */
        if(!empty($errors))
        {
            $c_row['errors'] = $errors;
        }   
            $output[$index][] = $c_row;
        }
        fclose($handle);
    }
    return $output;
}

$output = csv2array_with_validation('test.csv');

// Write error.json
if (isset($output['error']) && !empty($output['error']))  
{
    file_put_contents('error.json', json_encode($output['error'], JSON_PRETTY_PRINT | JSON_NUMERIC_CHECK));
}

// Write errorless.json
if (isset($output['error_less']) && !empty($output['error_less'])) 
{
    file_put_contents('error_less.json', json_encode($output['error_less'], JSON_PRETTY_PRINT | JSON_NUMERIC_CHECK));
}
?>

输出

$ php test.php
$ cat error.json 
[
    {
        "name": "BCD",
        "mark": -2,
        "url": "http:\/\/www.example.com",
        "errors": {
            "mark": "Mark should be between 0 to 5"
        }
    },
    {
        "name": "CD",
        "mark": 4,
        "url": "htt:\/\/www.c.com",
        "errors": {
            "url": "URL should be valid"
        }
    }
]
$ cat error_less.json 
[
    {
        "name": "ABCD",
        "mark": 5,
        "url": "http:\/\/www.example.org"
    }
]