CSV to Objects

Library to import CSV or XLSX files into an array of objects according to an specified definition

Instructions

To convert a file into an an array of objects, you must provide an ImportDefinition. This object is created by passing an array in an specific format. We recommend using YAML for defining this file as it is human-friendly. This format can be quite sophisticated, so two examples are provided:

Basic import definition

name: Fruits definition                                  # (Optional) Name of the import definition
columns:                                                 # (Required) An associative array with the headings of the columns in the file that will be imported
    Name:   { fruit: '#Name#' }                          # 'fruit' is a shortname for the object being created for the list (defined below) and #Name# will be the argument passed to the constructor. The hashes indicate that it should replace it with the value on that column                                         
    Weight: ~                                            # Null indicates that it can be ignored
classes:                                                 # (Required) An associative array with the mapped classes 
    fruit: 'CSVObjects\ImportBundle\Tests\Objects\Fruit' # The first one will be the class that will be returned 

Full import definition

All the extra properties are optional.

name: Fruits definition
columns:
    File Version:  { expect: 38 }                                                                                             # Expects indicates that the row value must always be as stated
    Name:          { fruit: ['#Name#', '#Colour#', '#Origin#', '#Origin - City#', '#Class#', '#Expiry date#', '#Contract#'] } # The constructor now takes an array of arguments
    Weight:        ~
    Colour:        { validate: ['red', 'yellow'] }                                                                            # Valid values that this column could have
    Origin:        { map: { 'Granada': 'Spain', 'Malaga': 'Spain', 'Dover': 'UK' } }                                          # It will apply this alias to the data on this column. If no validate is provided, non-matching values will convert into null
    Origin - City: ~
    Class:         { extract: 'Class (.*)', validate: ['A+', 'A', 'B', 'C'] }                                                 # It is like mapping but using the expecified regular expression to do the conversion. The desired part must be in brackets. Functions can optionally be combined when required
    Expiry date:   { validate: 'date', sourceFormat: 'd/m/Y', format: 'Y-m-d' }                                               # If validate is not an array, it can be an special value. 'date' will check that it is a date. Optionally, you can specify the source format using 'sourceFormat' or reformat it by specifying 'format'
    Contract:      { contract: ['#Contract#'] }
copy:                                                                                                                         # If specified, if will add more columns to the row, copying from the referenced columns before being processed
    Origin - City: Origin
classes:
    fruit:    ['CSVObjects\ImportBundle\Tests\Objects\Fruit', 'getFruitFromFullInfo']                                         # If the mapped class is an array, it will make a static call to the specified method to create the class
    contract: ['CSVObjects\ImportBundle\Tests\Objects\Contract', 'getContractFromId' ]

For the object mappings, you can also use a service in the ['@service_name', 'method'] notation.

For a more complex example see school-results.yml and school-results.csv.