This page is for a data modeller who needs to load data from an external source via HTTP.

Overview

An HTTP collector reads data from a HTTP Datasource. The collector defines how the data needed from the datasource is extracted to be used in PhixFlow.

To add a new HTTP collector to an analysis model:

To add an existing HTTP collector to an analysis model, in the model diagram toolbar:

Table Values in a HTTP Collector

To drive the lookups made by a HTTP collector from a table, the two must be connected using a lookup pipe. For example, a URL for a server can either be captured or calculated in an attribute called "ServerURL". The URL is then passed via a lookup pike to the HTTP collector to be used in its URL Expression. 

If the pipe is called in, here is how the URL Expression would be written on the HTTP collector: 
  {substring(in.ServerUrl,9)}

Basic Settings

The datasource to which this collector is connected may list multiple instances from which the data may be accessed. Each HTTP datasource instance is identified by a unique string. This expression should evaluate to a string which allows the collector to determine the specific instance to use. If the expression is blank then the collector will assume that there is only one instance and will use that one by default. If there is more than one instance and no expression is provided here then an error will be thrown during analysis since the collector will be unable to determine which source to use.

See also  HTTP datasource instance and Expressions and PhixScripts.

Send Message

Define details of the HTTP request sent to the HTTP Datasource to get the data required.

The URL to be used, without the leading http:// prefix. The URL may contain embedded expressions within { }. If this field is blank, the URL field on the httpDatasourceInstance is used directly.

For Example, this expression adds to the base URL provided by the HTTP datasource instance:
    {_url}/sub1/sub2?param1=3

The HTTP Collector will follow any HTTP redirections and return the final response.

See also HTTP datasource instance and Expressions and PhixScripts.

<?xml version ="1.0"?> <!DOCTYPE CORPORATE DASHBOARD "corpDash.dtd"> <results user="%USERNAME%" password="%PASSWORD%"> <monthlyTotals region={'"' + Region + '"'} division={'"' + Division + '"'}> <totalBilled>{'"' + TotalBilled + '"'}<\totalBilled> <totalCollected>{'"' + TotalCollected + '"'}<\totalCollected> <monthlyTotals> <\results>

The username and password for the HTTP Datasource Instance are available as %USERNAME% and %PASSWORD%.

The data will be encoded using the charset parameter specified by the Content-Type Header if one is present. If no Content-Type Header is set then ISO-8859-1 will be used. If the Content-Type header is set, but does not specify a charset then PhixFlow will use a default character set dependant on the content type.

HTTP Headers

Some headers will be set to default values if not provided. Automatically added headers may not appear in the debug log. The Content-Length header will be added to all requests and cannot be overridden by providing a value.

Response

Define the data response type/format that will be returned:

• HTML: response type allows an XPath Expression to be specified in order to retrieve just specified sections of the data into XML structures.
• XML: response type allows an XPath Expression to be specified in order to retrieve just specified sections of the data into XML structures. XML response types also support XML namespaces. The Xml Namspaces tab will be available when this response type is chosen.
• String: response type will return the full data as a string value.

Please see Response Examples for how the returned data can be used and evaluated in the corresponding stream attribute expressions.

The XPath expression used to resolve or filter the data that comes back in XML or HTML format.

XML Namespaces

This section has a toolbar with standard buttons. The grid contains a list of the namespaces defined in an XML response.

Advanced

XML Data

<?xml version ="1.0"?> <root xmlns:h="http://www.w3.org/TR/html4/"> <main page="PF Main Page" > <h:title h:name="PF Title">PF Title Text <h:datarow> <h:data h:initials="AA">Alistair Andrews</h:data> <h:data h:initials="BB">Bert Brown</h:data> </h:datarow> </h:title> <title name="Non namespace Title">Non namespace Title Text</title> </main> </root>

HTML Data

<html> <body nodename="Html Body"> <table> <tbody> <tr> <td initials="AA">Alistair Andrews</td> <td initials="BB">Bert Brown</td> </tr> </tbody> </table> </body> </html>

The data is being pointed to by either HTTP datasources or XML/HTML File collectors respectively.

The following table shows the different types of responses that can be returned from an HTTP Collector and how these can be used in the corresponding stream attribute expressions. A HTTP Collector response type of XML/HTML will mimic the responses from XML/HTML Collectors respectively.

<?xml version ="1.0"?>
<!DOCTYPE CORPORATE DASHBOARD "corpDash.dtd">
<results user="%USERNAME%" password="%PASSWORD%">
	<monthlyTotals region=${'"' + Region + '"'} division=${'"' + Division + '"'}>
		<totalBilled>${'"' + TotalBilled + '"'}</totalBilled>
		<totalCollected>${'"' + TotalCollected + '"'}</totalCollected>
	</monthlyTotals>
</results>

{
	user: '${user}',
	code:  ${'{size: "big"}'}
	price: ${price},
	currency: '${currency}'
}

<root> 
	<main page="PF Main Page"> 
		<title name="PF Title">PF Title Text
			<datarow> 
				<data initials="AA">Alistair Andrews</data>
				<data initials="BB">Bert Brown</data> 
			</datarow> 
		</title> 
	</main> 
</root>

<root xmlns:h="http://example.com/schema"> 
	<main page="PF Main Page"> 
		<h:title name="PF Title">PF Title Text
			<h:datarow> 
				<h:data h:initials="AA">Alistair Andrews</data>
				<h:data h:initials="BB">Bert Brown</data> 
			</h:datarow> 
		</h:title> 
	</main> 
</root>

<html>
	<body nodename="Html Body"> 
		<table> 
			<tbody> 
				<tr> 
					<td initials="AA">Alistair Andrews</td> 
					<td initials="BB">Bert Brown</td> 
				</tr> 
			</tbody> 
		</table> 
	</body> 
</html>

{
	"main_page": {
		"page": "PF Main Page",
		"title" : {
			"name" : "PF Title Text,
			"data" : [
				{"initials": "AA", "value" : "Alistair Andrews"},
				{"initials": "BB", "value" : "Bert Brown"}
			]
		}
	}
}

/html/body/table

Note the use of:

• ^ to traverse to the immediate parent element.
• the listToString function to handle multiple matching child elements/attributes.
• the optional html <tbody> tags. If these are not in your html data, then PhixFlow will insert them to conform with the HTML standards. Therefore when using absolute XPath expressions, the tbody tags need to be included even if they are not present in the incoming HTML data. That means /html/title/table/tr should be replaced with /html/title/table/tbody/tr or use //tr
  The same applies when referencing parent/child nodes within the stream attribute expressions.

Advanced

• Log HTTP Collector Connection Details : when ticked, PhixFlow always logs HTTP responses and requests for HTTP collectors, whatever is set here.