Have you ever wanted to have something like an ORM but for CSV files? No? Well now you can!
Introducing CSV Schema Parser. The number one way to turn your boring old CSV files into kick-ass PHP objects. And as if that wasn't amazing enough, it also casts your data to the correct data type! How cool is that? Not very cool you say? Well I disagree!
composer require sassnowski/csv-schema
First we have to define the schema of the CSV file we're about to parse. The schema is an associative array where the keys define what properties will be named on the resulting object. The values specify the data type of that column. The schema is ordered, meaning the first entry in the schema will correspond to the first column in the CSV and so on.
<?php
$config = [
'schema' => [
'first_name' => 'string',
'last_name' => 'string',
'age' => 'integer',
'coolness_factor' => 'float',
]
];
After we defined the schema, we can instantiate a new parser from it.
<?php
// continuing from above..
$parser = new Sassnowski\CsvSchema\Parser($config);
The parser provides a fromString
method that will turn any CSV string into an array of objects. The objects will be structured according to our schema.
<?php
public Parser::fromString(string $input): array
<?php
// Using our parser from above..
$input = "Kai,Sassnowski,26,0.3\nJohn,Doe,38,7.8";
$rows = $parser->fromString($input);
var_dump($rows[0]->firstname);
// string(3) "Kai"
var_dump($rows[0]->age);
// int(26)
var_dump($rows[0]->coolness_factor);
// float(0.3)
More often than not you will parse your CSV from a file. For this the Parser
provides the fromFile
method.
<?php
public Parser::fromFile(string $filename): array
<?php
// Using our parser from above..
// Assuming our file contains the same data as the string example.
$rows = $parser->fromFile('somefile.csv');
var_dump($rows[1]->firstname);
// string(4) "John"
var_dump($rows[1]->coolness_factor);
// float(7.8)
The configuration array provides a way to overwrite the default settings of the parser. You can set the delimiter
, the enclosure
character, the escape
character and the encoding
of the input file.
<?php
$config = [
'delimiter' => ',',
'enclosure' => '"',
'escape' => '\\',
'skipTitle' => false,
'encoding' => 'UTF-8',
'schema' => [ /* Your schema */ ],
];
You can parse a column to string
, float
, integer
and array
.
Assuming you have multiple values in a column (sometimes we cannot choose our data...) you might want to parse that column into an array instead.
You can do this by specifying array:<delimiter>
as the column type in your schema.
<?php
$config = [
'schema' => [
'name' => 'string',
'friends' => 'array:|'
],
];
$input = 'Kai Sassnowski,Lots|and|lots|of|friends';
$parser = new Sassnowski\CsvSchema\Parser($config);
$rows = $parser->fromString($input);
var_dump($rows[0])->friends);
// array(5) {
// [0]=>
// string(4) "Lots"
// [1]=>
// string(3) "and"
// [2]=>
// string(4) "lots"
// [3]=>
// string(2) "of"
// [4]=>
// string(7) "friends"
// }
Sometimes you might want a bit more control than the built-in types give you. For instance, you might want to query your database
based on an integer in one of your columns and return a model instead. You can easily register
arbitrarily complex types by using the static registerType
method on the Parser
class.
<?php
public static registerType(string $type, callable $callback)
The provided callback receives the column's value and should return its parsed representation.
<?php
Parser::registerType('foo', function ($value) {
return $value.'foo';
});
$config = [
'schema' => [
'custom' => 'foo'
]
];
$parser = new \Sassnowski\CsvSchema\Parser($config);
$rows = $parser->fromString("hello\nworld");
var_dump($rows[0]->custom);
// string(8) "hellofoo"
var_dump($rows[1]->custom);
// string(8) "worldfoo"
You can add additional parameters to your types by specifying them in the following format in your schema <type>:<parameter>
.
In this case, the callback function gets passed a second parameter containing the parameter specified in the schema.
This allows you to reuse your types instead of defining all sorts of variations of the same type (like querying the database, but using a different table/model).
<?php
Parser::registerType('model', function ($value, $table) {
// Pseudo code, assuming some library.
return DB::table($table)->findById($value);
});
$config = [
'schema' => [
'author' => 'model:authors',
'uploaded_by' => 'model:users',
]
];
$input = "5,13";
$parser = new \Sassnowski\CsvSchema\Parser($config);
$rows = $parser->fromString($input);
var_dump($rows[0]->author);
// object(Author)#1 (15) {
// ["id"]=>
// int(5)
// ...
// }
var_dump($rows[1]->uploaded_by);
// object(User)#1 (12) {
// ["id"]=>
// int(13)
// ...
// }