<?php
require('../vendor/autoload.php');

// Import the classes we need
use EnvoyMediaGroup\Columna\Writer;
use EnvoyMediaGroup\Columna\Reader;
use EnvoyMediaGroup\Columna\ColumnDefinition;

// Create or retrieve our data set
$array = [
    [
        'clicks' => 12,
        'platform_id' => 2,
        'site_id' => 7,
        'url' => 'https://www.foo.com',
    ],
    [
        'clicks' => 31,
        'platform_id' => 2,
        'site_id' => 9,
        'url' => 'https://www.barbaz.net',
    ],
    //... etc.
];

// Define our Metric and our Dimensions
// The names should match the keys in your data set
$MetricDefinition = new ColumnDefinition(
    ColumnDefinition::AXIS_TYPE_METRIC, // metric or dimension
    'clicks',                           // name (should match the keys in your data set)
    ColumnDefinition::DATA_TYPE_INT,    // data type (string, int, float, bool, datetime)
    null,                               // precision (for floats)
    0                                   // empty value (matching the specified data type)
);

$DimensionDefinitions = [
    new ColumnDefinition(
        ColumnDefinition::AXIS_TYPE_DIMENSION,
        'platform_id',
        ColumnDefinition::DATA_TYPE_INT,
        null,
        0
    ),
    new ColumnDefinition(
        ColumnDefinition::AXIS_TYPE_DIMENSION,
        'site_id',
        ColumnDefinition::DATA_TYPE_INT,
        null,
        0
    ),
    new ColumnDefinition(
        ColumnDefinition::AXIS_TYPE_DIMENSION,
        'url',
        ColumnDefinition::DATA_TYPE_STRING,
        null,
        ''
    ),
];

// Set our output path and the date for this file's data
$date = '2022-07-08';
$file_path = "/data_directory/{$date}/{$MetricDefinition->getName()}." . Reader::FILE_EXTENSION;

// Instantiate the Writer 
$Writer = new Writer();

// The Writer expects headers (string keys) separate from data (0-indexed).
//   If your data is associative like the above, you can separate it with
//   this helper function.
list($headers,$data) = $Writer->separateHeadersAndData($array);

// Write the columnar file
$Writer->writeFile(
    $date,
    $MetricDefinition,
    $DimensionDefinitions,
    $headers,
    $data,
    $file_path,
    // Some optional flags:
    //$do_rle_compression = true,   // Perform run-length encoding (RLE) compression
    //$do_cardinality_sort = false, // Sort data by cardinality of columns before RLE
    //$lock_output_file = true      // Acquire an exclusive lock when writing output file
);

<?php
require('../vendor/autoload.php');

// Import classes
use EnvoyMediaGroup\Columna\CombinedWriter;
use EnvoyMediaGroup\Columna\ColumnDefinition;

// Set our arguments
$date = '2022-07-08';
$metric = 'clicks';
$partial_files = [
    "/tmp/{$date}/{$metric}/partial_1.scf",
    "/tmp/{$date}/{$metric}/partial_2.scf",
    "/tmp/{$date}/{$metric}/partial_3.scf",
    //... etc.
];
$combined_file_path = "/data_directory/{$date}/{$metric}." . Reader::FILE_EXTENSION;

// See Writer example above for metric and dimension definitions
$MetricDefinition = new ColumnDefinition(...); 
$DimensionDefinitions = [...];

// Write the combined file from the partial files
$Writer = new CombinedWriter();
$response = $Writer->writeCombinedFile(
    $date,
    $MetricDefinition,
    $DimensionDefinitions,
    $partial_files,
    $combined_file_path
    // Some optional flags:
    //$lock_output_file = true      // Acquire an exclusive lock when writing output file
);

// If you want to remove the partial files after creating the combined file:
foreach ($partial_files as $partial_file) {
    unlink($partial_file);
}

<?php

// Import the needed classes
use EnvoyMediaGroup\Columna\BundledReader as Reader;
use EnvoyMediaGroup\Columna\Constraint;

// Specify our metric and date, and the corresponding file path
$metric = 'clicks';
$date = '2022-07-08';
$file_path = "/data_directory/{$date}/{$metric}." . Reader::FILE_EXTENSION;

// Set what dimensions we want to include in our results
$dimensions = [
    'platform_id',
    'site_id',
];

// Define our constraints
// Constraints are ANDed within groups, ORed between groups
// This example is equivalent to the following SQL:
//   SELECT * FROM file WHERE (platform_id = 7 AND site_id in (1,3,17)) OR (url LIKE '%sale%');
$constraints = [
    [
        (new Constraint("platform_id",Constraint::EQUALS,7))->toArray(),
        (new Constraint("site_id",Constraint::IN,[1,3,17]))->toArray(),
    ],
    [
        (new Constraint("url",Constraint::CONTAINS,"sale"))->toArray(),
    ]
];

// Group the results by the dimensions we asked for (in this case, platform_id and site_id)
$do_aggregate = true;
// Don't provide extra metadata with sum/count/min/max for each grouping, just aggregate the values
$do_aggregate_meta = false;

$Reader = new Reader();
$Reader->run(
    $date,
    $metric,
    $dimensions,
    $constraints,
    $do_aggregate,
    $do_aggregate_meta,
    $file_path
);

$metadata = $Reader->getMetadata(); // Metadata about the request and results; see sample below.
$data = $Reader->getResults(); // Results of the request; see sample below.

<?php

use EnvoyMediaGroup\Columna\Response;

// Craft your request
$workload_array = [
    "date" => "2022-07-08",
    "metric" => "clicks",
    "dimensions" => ["platform_id","site_id"],
    "constraints" => [
        [
            [
                "name" => "platform_id",
                "comparator" => ">=",
                "value" => 5,
            ],
        ],
    ],
    "do_aggregate" => true,
    "do_aggregate_meta" => false,
    "file" => "path/to/file.scf",
];
$workload = json_encode($workload_array);

// Transmit that workload over a network with your RPC framework of choice...
$result_string = $SomeRpcClient->request($workload);

// Unserialize the result with the Response class
$Response = new Response($result_string);

$metadata = $Response->getMetadata();
$results  = $Response->getResults();

<?php

use EnvoyMediaGroup\Columna\BundledReader as Reader;

$workload = $SomeRpcClient->receive();

$Reader = new Reader();
$Reader->runFromWorkload($workload);
$result_string = $Reader->getResponsePayload();

// Return that result string over your RPC framework...
$SomeRpcClient->respond($result_string);

Array(
  'date' => '2022-07-08', // Date of the file
  'metric' => 'clicks',   // Name of the metric in the file
  'status' => 'success',  // 'success' if records were found, 'empty' if no records were found, 'error' on failure
  'min' => 1,   // Least metric value among the records
  'max' => 64,  // Greatest metric value among the records
  'sum' => 102, // Total metric value among the records
  'matched_row_count' => 102, // Number of records in the file that matched your constraints prior to aggregation
  'result_row_count' => 17,   // Number of records in the result set after aggregation
  'column_meta' => Array( // Description of the columns in the result set
    0 => Array(
      // MD5 is automatically prepended. Records with matching dimension values will have matching md5 hashes.
      // This is helpful if you need to aggregate multiple Reader results together.
      'definition' => Array(
        'axis_type' => 'dimension',
        'name' => 'md5',
        'data_type' => 'string',
        'empty_value' => '',
      ),
      'index' => 0, // Numerical index in each result record that corresponds to this column
    ),
    1 => Array(
      'definition' => Array(
        'axis_type' => 'metric',
        'name' => 'clicks',
        'data_type' => 'int',
        'precision' => NULL,
        'empty_value' => 0,
      ),
      'index' => 1,
    ),
    2 => Array(
      'definition' => Array(
        'axis_type' => 'dimension',
        'name' => 'platform_id',
        'data_type' => 'int',
        'precision' => NULL,
        'empty_value' => 0,
      ),
      'index' => 2,
    ),
    3 => Array(
      'definition' => Array(
        'axis_type' => 'dimension',
        'name' => 'site_id',
        'data_type' => 'int',
        'precision' => NULL,
        'empty_value' => 0,
      ),
      'index' => 3,
    ),
  ),
  'is_aggregated' => true, // Read back of whether these results are aggregated on matching dimension values.
  'aggregate_includes_meta' => false, // Read back of whether the results include metadata for each aggregate grouping.
  'host' => 'worker-1', // Result of php_uname('n'), helpful for tracing/debugging
  'ms_elapsed' => 32,   // Milliseconds it took to complete your request
)

Array(
    0 => Array (
        0 => 'a060e57689d68664f873561a78e002d9',
        1 => 3,
        2 => 58,
        3 => 1,
    ),
    1 => Array (
        0 => 'f9f0a70e4d259b63914ccc98ed438d0e',
        1 => 16,
        2 => 54,
        3 => 1,
    ),
    2 => Array (
        0 => '166dea5e7a502516e662d389239bd2fc',
        1 => 4,
        2 => 75,
        3 => 1,
    ),
    // ... etc.
)