使用SimpleXML将多个XML文件转换为一个CSV

Question

I have some xml files, which have the same elements but only with different information.

First file test.xml

<?xml version="1.0" encoding="UTF-8"?>
<phones>
    <phone>
        <title>"Apple iPhone 5S"</title>
        <price>
            <regularprice>500</regularprice>
            <saleprice>480</saleprice>
        </price> 
        <color>black</color>
    </phone>
</phones>

Second file test1.xml

<?xml version="1.0" encoding="UTF-8"?>
<phones>
    <phone>
        <title>Nokia Lumia 830</title>
        <price>
            <regularprice>400</regularprice>
            <saleprice>370</saleprice>
        </price> 
        <color>black</color>
    </phone>
</phones>

I need to convert some values from these xml files into 1 test.csv file

So I am using this php code

<?php

$filexml1='test.xml';
$filexml2='test1.xml';

    //File 1
    if (file_exists($filexml1)) {
        $xml = simplexml_load_file($filexml1); 
        $f = fopen('test.csv', 'w');

    $headers = array('title', 'color');
    $converted_array = array_map("strtoupper", $headers);


    fputcsv($f, $converted_array, ',', '"');


    foreach ($xml->phone as $phone) {

        //$phone->title = trim($phone->title, " ");
        // Array of just the components you need...
        $values = array(
           "title" => (string)$phone->title = trim(str_replace ( "\"", "&quot;", $phone->title ), " "), 
           "color" => (string)$phone->color
        );
        fputcsv($f, $values,',','"');

    }
    fclose($f); 

    echo "<p>File 1 coverted to .csv sucessfully</p>";
} else {
    exit('Failed to open test.xml.');
}

    //File 2
    if (file_exists($filexml2)) {
        $xml = simplexml_load_file($filexml2); 
        $f = fopen('test.csv', 'a');


    //the same code for second file like for the first file

    echo "<p>File 2 coverted to .csv sucessfully</p>";
} else {
    exit('Failed to open test1.xml.');
}

?>

The output of the test.csv looks this way

TITLE             COLOR
Apple iPhone 5S   black
Nokia Lumia 830   black

As you can see I only managed to load each file into a variable and for each file I have to write if statement which makes the script too big, so I am wondering if it is possible to load all files into array, process them with one code block because xml elements are the same and output to one .csv file? Essentially I need the same test.csv output only with less php code.

Thanks in advance.

Answer 1

Next to using an array, there is more in PHP which can make it even more simple. Like an array could represent a list of your files, other constructs in PHP can that, too.

For example, as the XML files you have most likely are inside a specific directory and follow some pattern with their filename, those could be easily represented with a GlobIterator :

$inputFiles = new GlobIterator(__DIR__ . '/*.xml');

You could then foreach over them which I'll show in a moment with another example.

Such a list allows you to streamline your processing. That is important because there is some kind of a generic formular for many programs: Input, Process, Output. This is also called IPO or IPO+S Model. The S stands for storing. In your case while you process the input data, you also store into a new file CSV file which is also the output (after processing is fully done).

When you follow such a generic model, it's easier to structure your code and with a better structure you most often have less code. Even if not, each part of your code is more self-contained and smaller which is most often what you're looking for.

Next to the said list of XML-files I showed at the beginning of the answer with the GlobIterator there are other Iterators that can help to process the XML data.

For example, you've got 1-n XML files that contain 0-n <phone> elements. You know that you want to process any of these <phone> elements, you already exactly know what you want to do with them (extract some data from it). So wouldn't it be great to have a list of all <phone> elements within all XML-files first?

This can be easily done in PHP with the help of a Generator . That is a function that can return values multiple times while it's still "running". This is a simplification, better show some code to illustrate that. Let's say we've got the list of XML files as input and we want all <phone> elements out of it. For sure, you could create an array of all these <phone> elements and process that array later. However, a Generator is able to offer all these <phone> elements directly to be used within a foreach loop:

function extract_phones(Traversable $files) {
    foreach ($files as $file) {
        $xml = simplexml_load_file($file);
        if ($xml === false) {
            continue;
        }
        foreach ($xml->phone as $phone) {
            yield $phone;
        }
    }
}

As this exemplary Generator function shows, it goes over all $files , tries to load them as a SimpleXMLElement and if successfull, iterates over all <phone> elements and yields them.

That means, if the function extract_phones is called within a foreach , that loop will have every <phone> element as SimpleXMLElement :

foreach(extract_phones($inputFiles) as $phone) {
    # $phone is a SimpleXMLElement here
}

So now your question asks about creating the CSV file as output. This could be done creating an SplFileObject to pass the output around and access it while processing. It basically works the same like passing the file-handle around like you do in your question but it has better semantics that do allow to change the code more easily later on (you could replace it with another object that behaves the same).

Additionally I've seen a little detail in your code that is worth for some discussion first. You're encoding the quotes as HTML entities:

 trim(str_replace( "\"", """, $phone->title ), " ")

You most likely do that because you want to have HTML-Entities inside the CSV file. However, the CSV file does not need such. You also want to have the data in the CSV file as generic as possible. Whether the CSV file is used inside a HTML context later on or within a spreadsheet application should not be your concern when you convert the file-format. My suggestion is here to leave that out and deal at another place with it. A place this more belongs to, and that is later on, eg if you use the data from the CSV creating some HTML.

That keeps your conversion and the data clean and it also removes detailed places in your processing which not only make the code more complicate but are very often a place where we introduce flaws into our programs.

I for myself will just remove it from my example.

So let's put this all together: Get all phones from all XML files and store the fields interested in into the output CSV file:

$files  = new GlobIterator(__DIR__ . '/*.xml');
$phones = extract_phones($files);

$output = new SplFileObject('file.csv', 'w');
$output->fputcsv($header = ["title", "color"]);

foreach ($phones as $phone) {
    $output->fputcsv(
        [
            $phone->title,
            $phone->color,
        ]
    );
}

This then creates the output file you're looking for (without the HTML-entities):

title,color
"""Apple iPhone 5S""",black
"Nokia Lumia 830",black

All this needs is the generator-function I've showed above already that in itself has also straight-forward code. Everything else ships with PHP already. Here is the example code in full:

<?php
/**
 * @link http://stackoverflow.com/questions/26074850/convert-multiple-xml-files-to-csv-with-simplexml
 */

function extract_phones(Traversable $files)
{
    foreach ($files as $file) {
        $xml = simplexml_load_file($file);
        if ($xml === false) {
            continue;
        }
        foreach ($xml->phone as $phone) {
            yield $phone;
        }
    }
}

$files  = new GlobIterator(__DIR__ . '/*.xml');
$phones = extract_phones($files);

$output = new SplFileObject('file.csv', 'w');
$output->fputcsv($header = ["title", "color"]);

foreach ($phones as $phone) {
    $output->fputcsv(
        [
            $phone->title,
            $phone->color,
        ]
    );
}

echo file_get_contents($output->getFilename());

Answer 2

Thanks @Ghost for pointing me to the right direction. So here is my solution.

<?php

$filexml = array ('test.xml', 'test1.xml');


//Headers
$fp = fopen('file.csv', 'w');

$headers = array('title', 'color');
$converted_array = array_map("strtoupper", $headers);


fputcsv($fp, $converted_array, ',', '"');


//XML
foreach ($filexml as $file) {
    if (file_exists($file)) {
        $xml = simplexml_load_file($file);

        foreach ($xml->phone as $phone) {
        $values = array(
               "title" => (string)$phone->title = trim(str_replace ( "\"", "&quot;", $phone->title ), " "), 
               "color" => (string)$phone->color
            );
            fputcsv($fp, $values, ',', '"');
        }
        echo $file . ' converted to .csv sucessfully' . '<br>';
    } else {
        echo $file . ' was not found' . '<br>';
    }


}

fclose($fp);

?>