CSV is a format that has been around for a long time. It's a simple way to store multiple rows with multiple columns and it can be easily imported into a variety of commonly used desktop applications. This article examines the use of CSV-formatted data in Flash. We will write a class in ActionScript 2.0 to handle all of our CSV functionality and then we will use this class to populate the Macromedia v2 Data Grid Component.

Why Use a CSV File?
CSV is a very simple format. At its core it is essentially a large string that has rows separated by a delimiter and columns within each row delimited by a comma. Because of its use of rows and columns, this data format is perfectly suited for record sets, whereas an XML-formatted record set is more complicated.

Another good reason to use CSV is that it can be viewed by many common desktop applications. Any spreadsheet program, such as Microsoft Excel, can view, modify, or create CSV files. Unlike an XML-formatted record set, users can open a CSV file and easily read the information.

Like XML, CSV is data formatted within a file. It's not tied to any specific software or platform and it doesn't need a server in order to deliver it to the client. This makes CSV and XML files a very portable way to handle data. You could build a Web- based Flash application using a CSV or XML file and then easily use that same code to build a CD-ROM-based application.

CSV will not always be the right solution for your project. XML is usually a better solution for delivering data to the client because of its versatility. Any time your data is not formatted in a series of rows and columns, CSV should be avoided. You should also refrain from using the CSV format if you have multiple levels of complex data types. For example, if you have an array or a list in one of your cells, you should probably use XML to define your data. XML is better at separating out different data types and handling multiple levels of embedded data. CSV should only be used for primitive data types (String, Number, and Boolean) that are housed in the row-and-column format.

CSV Specifications
The first thing you'll notice about a CSV file is that it has rows and columns. The rows are delimited by a line break, a carriage return, or a combination of the two. The columns within each row are separated by a comma. Many times the first row of the file is used to define the column headers; we'll talk more about that when we begin parsing the data in our custom class.

Many CSV files will use something called a qualifier, the most common qualifier being the double quote ("). This character is used to surround each column's data. Below you can see the difference between a file with a qualifier and one without a qualifier.

CSV data without a qualifier:
Field 1,Field 2,Field 3,Field 4
Field 5,Field 6,Field 7,Field 8

CSV data with a qualifier:
"Field 1","Field 2","Field 3","Field 4"
"Field 5","Field 6","Field 7","Field 8"

Building the CSV Class
Before we begin programming our CSV class, we need to define exactly what we need this class to do:

1. The class will need to load in an external CSV file.
2. The class will need to parse the data into a Flash object. The format we want to get the CSV data into would be an array of objects (or associative arrays), with the object's key being the column header.
3. The parsing will need to be able to pull the column headers from the first row of the CSV; however, we will also want to give the user the option to override this feature by specifically defining the columns in an array prior to parsing the file.
4. We also need the parser to be able to handle qualifiers. We will allow the user to define the qualifier prior to parsing the CSV data.

class CSV {
	function CSV() {}
}

Next we need to define our class's parameters:

• csvData is a private array that will hold the parsed CSV data.
• onLoad is a public method that the user can overwrite to get a callback when the data is loaded and parsed.
• columns is a public array that the user can define prior to parsing the CSV data to manually define the column headers.
• qualifier is a public string that the user can define prior to parsing the CSV data. Qualifiers assist with more accurate parsing. We will default this value to an empty string.

private var csvData:Array;
public var onLoad:Function;
public var columns:Array;
public var qualifier:String = '';

The first task that we need our class to do is load data in from a file. We'll accomplish this by creating a load method, which will simply take in the CSV file's path as its only parameter. The load method will use the LoadVars class to load the file into Flash, but we will actually incorporate an undocumented method within the LoadVars class, called onData. This method works exactly like the XML.onData method. It allows us to get the raw file data without having it parsed by the LoadVars class.

public function load(csvPath:String):Void {
	var csvLoad:LoadVars = new LoadVars();
	csvLoad._parent = this;
	csvLoad.onData = function(rawData:String) {
		this._parent.onData(rawData);
	}
	csvLoad.load(csvPath);
}

Notice that in our LoadVars.onData method we are calling a method within our class, called onData. We will create that method to handle the parsing of the CSV data after the file has loaded. Once the data is parsed it will pass it back to the user through the onLoad method. We won't actually define the onLoad method. To handle the load method callback, the user will define this method. Users could also override the onData method if they didn't want the data parsed right away.

public function onData(rawData:String):Void {
	csvData = parseCSV(rawData);
	onLoad();
}

Next we'll add the parser method. We'll make this method public so users can utilize it without loading in the CSV data from a file. The parser is the most complicated part of this class - it is also the part of the class where the most can go wrong.

Our parser method will take a raw CSV string as its sole parameter and return the parsed data as an array of objects. The first two things our parser method must do is determine the row and column delimiters. We already know that the row delimiter is a carriage return (\r), a new line (\n), or both (\r\n). A quick search of the string will help us determine what the actual row delimiter is. The column delimiter is always a comma in the CSV specification; however, we can make our parsing more accurate by adding the qualifier before and after the delimiter. This new concatenated string will be our column delimiter.

To begin parsing the CSV string we will split the string into an array of its rows. This will allow us to loop through the rows and parse each row individually.

If the column's parameter has not been defined prior to running this parser method, then we can assume the first row in the CSV file holds the column headings. We will populate the column's array from this row.

As we loop through the rows we will split each row into an array of its columns. If the row doesn't have the same number of columns as the columns parameter, then we will skip that row. As we loop through the columns within each row we will place the data into an object using the column heading as the key. We will then append this object to the end of the array that we will be returning (see Code I).

We also need to create a private method called removeQualifier that supports the parser. This method will pull out the qualifier at the beginning and end of each row (see Code II).

In order to expose the private csvData parameter, we will create a getter method called data. This method will simply return the parsed data to the user. We do this to protect the csvData parameter from being set outside the class.

public function get data():Array {
	return csvData;
}

Using the CSV Class with the Data Grid Component
In this example we will use Macromedia's Data Grid Component to display our CSV data. The data grid's dataProvider parameter accepts an array of objects as a valid value, so we won't have to manually populate the grid. The first thing we need to do is place an instance of the data grid on the stage, then give it an instance name of dataGrid_mc. Then we will create a new layer and name it Actions. We will put all of our code in frame 1 of the Actions layer. Make sure your CSV.as class file is in the same folder as your FLA file.

In frame 1 of our Actions layer we will begin writing our code, creating an instance of the CSV class.

var csvLoad:CSV = new CSV();

Next we define our columns and qualifier parameters before we parse the CSV file.

csvLoad.columns = new Array('Column 1', 'Column 2', 'Column 3');
csvLoad.qualifier = '"';

We then define our onLoad method, where we populate the data grid with the parsed CSV data. Finally, we call the load method and send it the path to our CSV file as an argument.

csvLoad.onLoad = function() {
	dataGrid_mc.dataProvider = this.data;
}
csvLoad.load('sampleData.csv');

Conclusion
I hope this article gives you an idea of how you can use CSV data within your Flash application. But more important, I hope it has given you a concrete example of how you can build an ActionScript 2.0 class and how you can make your class easy and flexible to the developers that implement it. This CSV class may not be the solution for your application, but the principles of object-oriented programming should be used in nearly every application you build with Flash.

Note: Class code can be downloaded from www.sys-con.com/mx/sourcec.cfm.