keboola-legacy/json-parser

Keboola JSON to CSV parser

Json Parser

Description

Parse Json strings into CSV files Creates multiple tables from a single JSON, if said JSON contains numbered arrays Uses Keboola\CsvFile for results, _(*1)

Usage

    use Keboola\Json\Parser;
    $parser = Parser::create(new \Monolog\Logger('json-parser'));
    $file = file_get_contents("some/data.json");
    $json = json_decode($file);

    $parser->process($json);

    $results = $parser->getCsvFiles(); // array of CsvFile objects

Parser\Json

Analyzes and parses JSON data into n*CSV files., _(*2)

create(\Monolog\Logger $logger, $struct, $analyzeRows)

• $struct should contain an array with results from previous analyze() calls (called automatically by process())
• $analyzeRows determines, how many rows of data (counting only the "root" level of each Json) will be analyzed [default -1 for infinite]

process($data, $type, $parentId)

• Expects an array of results as the $data parameter
• $type is used for naming the resulting table(s)
• The $parentId may be either a string, which will be saved in a JSON_parentId column, or an array with "column_name" => "value", which will name the column(s) by array key provided
• Checks whether the data needs to be analyzed, and either analyzes or parses it into $this->tables[$type] ($type is polished to comply with SAPI naming requirements)
• If the data is analyzed, it is stored in Cache and NOT PARSED until $this->getCsvFiles() is called

getCsvFiles()

• returns a list of \Common\Table objects with parse results

Parse characteristics

The analyze function loops through each row of an array (generally an array of results) and passes the row into analyzeRow() method. If the row only contains a string, it's stored in a "data" column, otherwise the row should usually be an object, so each of the object's variables will be used as a column name, and it's value analysed: - if it's a scalar, it'll be saved as a value of that column. - if it's another object, it'll be parsed recursively to analyzeRow(), with it's variable names prepended by current object's name - example: "parent": { "child" : "value1" } will result into a "parent_child" column with a string type of "value1" - if it's an array, it'll be passed to analyze() to create a new table, linked by JSON_parentId, _(*3)