Simple filter implementation for Pomm Project
Asking a collection of resources is often given with filter needs. But it's not always obvious to handle filters and build the query accordingly: multiple filters on different tables, multiple values for a filter, null values, dates, ...
This library provides a simple implementation to build query's condition from an array of filters.
- php >=5.4
composer require nmarniesse/pomm-filterThe library helps to create an instance of PommProject\Foundation\Where that you could use in every pomm query
(see here
for further explanation).
To explain what we can do with this library, we can take a practical case: we want to filter on active products with color 'blue' or 'yellow', in category 'accessory', with price between 50 and 100, and which have one tag.
The filter array representation is:
use NMarniesse\PommFilter\FilterInterface;
$array_filters = [
'is_active' => '1',
'color' => ['blue', 'yellow'],
'category' => ['accessory'],
'price_from' => 50,
'price_to' => 100,
'tag' => FilterInterface::_not_null_,
];With an HTTP query similar to ?filter[is_active]=1&filter[color][]=blue&filter[color][]=yellow&filter[category]=accessory&filter[price_from]=50&filter[price_from]=50&filter[price_to]=100&filter[tag]=_not_null_
you can have he same array in php with $array_filters = $_GET['filter'];.
You have your array filters, now let build the query:
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\BasicFilter;
use NMarniesse\PommFilter\FilterType\BooleanFilter;
# The sql query with a placeholder for the where condition
$sql = <<<SQL
SELECT
p.id,
p.color,
c.category_id,
pr.unit_price
FROM product p
INNER JOIN category c ON ...
INNER JOIN price pr ON ...
LEFT JOIN product_tag pt ON ...
WHERE {conditions}
SQL;
# Define the available filters and create the Where instance
$filter_condition = new FilterCondition('p');
$filter_condition->addFilter(new BasicFilter('color', 'p')); // optional
$filter_condition->addFilter(new BooleanFilter('is_active'));
$filter_condition->addFilter(new BasicFilter('category_id', 'c'));
$filter_condition->addFilter(new BasicFilter('unit_price', 'pr', '>='));
$filter_condition->addFilter(new BasicFilter('unit_price', 'pr', '<='), 'price_from');
$filter_condition->addFilter(new BasicFilter('tag', 'pt'), 'price_to');
// ...
$where = $filter_condition>getWhere($array_filters);
# Execute the query with Pomm with our instance of Where
$sql = str_replace('{conditions}', (string) $where, $sql);
$pomm_session->getQueryManager()->query($sql, $where->getValues());Important note
Even if the generated Where condition protects the query against SQL injection, please note you must clean and validate the data coming from users, according to your business rules.
By default the FilterCollection does not contain any filter.
The method getWhere($filters) convert any associative array into Where instance
When you do a getWhere(['key1' => 'val1']), it assumes that the key1 field exists in your query and
build a simple condition query key1 = $* with parameter 'val1'.
If you want to specify a table alias in the condition query, or not use the = operator, you have to
add the filter manually using the addFilter method.
Examples:
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\BasicFilter;
# Create a filter condition.
# When you pass a filter {"key1": "value1"}, it assumes that the field *key1* exists in your query
$filter_condition = new FilterCondition();
# When you have multiple tables in your query, you may specify the table/alias name
# Then when you pass a filter {"key1": "value1"}, it will automatically construct "user.key1 = $*"
$filter_condition = new FilterCondition('user');
# To use a filter on a field which is not on main table, you have to add it manually
# For example to add a filter on the field category on table p
$filter_condition->addFilter(new BasicFilter('category', 'p'));
# If you want personnalize your filter name, use second parameter to specify it
$filter_condition->addFilter(new BasicFilter('category', 'p'), 'my_custom_category_filter_name');This library provides several filter types to help you to create your own filter collection.
As its name indicates, this class is useful to create simple filter.
However you can specify the operator you want to use (default is =) in order to
customize the behavior of your filter.
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\BasicFilter;
# Create a BasicFilter
$filter1 = new BasicFilter('color');
# Create a BasicFilter and specify the table name/alias used in the query
$filter2 = new BasicFilter('category_id', 'c');
# If you want to filter on prices greater than specific value
$filter3 = new BasicFilter('unit_price', 'p', '>=');
# If you want to filter on prices greater than specific value
$filter4 = new BasicFilter('unit_price', 'p', '<=');
$filter_condition = new FilterCondition();
$filter_condition->addFilter($filter1);
$filter_condition->addFilter($filter2);
$filter_condition->addFilter($filter3, 'price_from');
$filter_condition->addFilter($filter4, 'price_to');
# Filter on color 'blue' or 'yellow', with category 'accessory', and price between 50 and 100
$filter_condition>getWhere([
'color' => ['blue', 'yellow'],
'category' => ['accessory'],
'price_from' => 50,
'price_to' => 100,
]);This filter allow you to use date values.
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\DateTimeFilter;
# Create DateTimeFilter
$filter1 = new DateTimeFilter('created_at', '', '>=');
$filter2 = new DateTimeFilter('created_at', '', '<=');
$filter_condition->addFilter($filter1, 'created_date_from');
$filter_condition->addFilter($filter2, 'created_date_to');
# Filter on color 'blue' or 'yellow', with category 'accessory', and price between 50 and 100
$filter_condition>getWhere([
'created_date_from' => '2010-01-01T00:00:00+00',
'created_date_to' => '2010-12-31T23:59:59+00',
]);This filter is used to handle boolean fields.
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\DateTimeFilter;
# Create BooleanFilter
$filter1 = new BooleanFilter('is_new');
$filter_condition->addFilter($filter1);
# Filter on true value
$filter_condition>getWhere([
'is_new' => true, // Any value different from false, 'inactive', 'false', '0', 0
]);
# Filter on true value
$filter_condition>getWhere([
'is_new' => false, // Any value among the values false, 'inactive', 'false', '0', 0
]);This filter is used to handle hstore fields.
Given we have a hstore field full_address which contains keys like street, city, postal code, country, etc...
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\HstoreFilter;
# Create HstoreFilter
$filter1 = new HstoreFilter('city', 'full_address');
$filter2 = new HstoreFilter('country_code', 'full_address');
$filter_condition->addFilter($filter1);
$filter_condition->addFilter($filter2);
# Filter on city value
$filter_condition>getWhere([
'city' => 'Paris',
]);
# Filter on country_code value
$filter_condition>getWhere([
'country_code' => 'FR',
]);This filter is used to handle ltree fields. As ltree is commonly used to handle tree views, you can filter on a value and all its descendants using this filter. If you don't want to filter on the descendants the BasicFilter is enough.
This filter is used to handle range fields. The value could be a single value or a range of values identified by a NMarniesse\PommFilter\ValueType\RangeValue object. In both cases, the filter tests the value is included into the range field.
use NMarniesse\PommFilter\FilterCondition;
use NMarniesse\PommFilter\FilterType\RangeFilter;
use NMarniesse\PommFilter\ValueType\RangeValue;
# Create a RangeFilter
$filter1 = new RangeFilter('score_range');
$filter2 = new RangeFilter('lifetime');
$filter_condition->addFilter($filter1);
$filter_condition->addFilter($filter2);
# Filter entities which have a score 10 and be active in 2010 january
$filter_condition>getWhere([
'score_range' => 10,
'lifetime' => new RangeValue(
new \DateTime('2010-01-01 00:00:00+00:00'),
new \DateTime('2010-12-31 23:59:59+00:00')
),
]);make unit-tests