Table Of Contents

Graph ► Tools ► Fetch

Use the Fetch tool to retrieve data from a web server or the file system on your computer as referenced by a URL. The Fetch tool is diagrammed here:

URL Types

There are two types of URLs that can be used as defined here:

• File: A file URL begins with "file://" and the rest of the URL is the path to the data file. Since paths are absolute there are three leading slashes, such as: "file:///Users/steve/data/mypoints.txt" (without the quotes). In this example, the path extension is txt however in practice it can be many things or the file can be without extension.

• Web: A web URL begins with "http://" and the rest of the URL is the path to the data file, such as: "http://www.vvi.com/data/mypoints.txt" (without the quotes). Notice that the path extension is "txt" in this case. Although the extension is arbitrary, txt is a safe extension as it informs a web server to use an ASCII MIME type. In practice, any extension, or no extension, also works. This is the same URL that you would type into a web browser to download the data. You can use the data directly in this manner, or first download it and then use the File type URL. Either way, the data file must be a resource of the web server that the URL points to, either as a static file or a dynamic URL that retrieves the data file bytes from other places.

Notice that the Web URL gives access to a programmable system. For example, you can turn on your Web Sharing in the system preferences and then use PHP, perl or other scripting languages to write algorithms to retrieve computed data. The URL would be of this form: "http://localhost/cgi-bin/myalgorithm?data=lab1&section=2&param1=5" and you use the scripting engine's built-in form facilities to parse the URL parameters as input to an algorithm. Alternatively, the URL could point to an existing web service to retrieve SOA type information from queries.

Content Format Type

If the component pop up button does not exist or a column number is selected on it then the content retrieved from a URL (the response bytes) is formatted according to the list below. Note that the fetch content format is simple, it is just a list of numbers with a blank delimiter. No fancy XML formatting and no dimension data. It is intended be as simple as possible in order to facilitate ease of use. Data formats are also explained in each Task's Help tool and section in this manual.

• A list of numbers for the Set Of Scalars task.

• A list of 2D points (two numbers) for the Set Of 2D Points task, such as: x1 y1 x2 y2 ... xN yN.

• A list of 3D points (3 numbers) for the 3D Points Task.

• A list of numbers for the Z Values and Density tasks.

If the component pop up button does not exist or "All" is selected then the content retrieved from a URL (the response bytes) can be XML. For an explanation of this content see the XML Fetch tutorial.

Fetch Timing

Fetching data can occur at different time intervals as defined here:

• Fetch Data: Clicking the Fetch Data button fetches the data immediately.

• Each Animation Step: Selecting the Each Animation Step switch causes the data to be fetched upon each animation, which is at an interval of one second. If you use this button then go to the Data tool and select Animate.

• At First Appearance: Selecting the At First Appearance switch causes the data to be fetched when the task is made visible. Tasks are only made visible when switching tasks so this operation does not occur too frequently.

Fetch Operation

In a nutshell, when a fetch occurs the data is retrieved, entered into the task and then displayed. Usually the fetch appears to be atomic, that is: it seems to be a single operation. However, fetching is asynchronous and non-atomic. You may notice this fact if the fetch takes a long time. Also, for tasks that have multiple columns the fetch is on a per-column basis and the fetch retrieves each column individually until all columns are fetched and then updates the task with the fetched data. Multiple column fetches are done in parallel and asynchronously and each column fetch rendezvous to amalgamate the fetch into a single result.