The Output Step
The "Processing" step is followed by the output step. In the output step, data is generated into the appropriate format, and transferred either synchronously to one target system or successively to several target systems. You can use content-related conditions to control whether only certain formats are allowed for a specific type of document, as well as whether only certain target systems are addressed.
How an output worker selects the available documents from the database is analogous to how a processing worker works (see The Processing Step). Output might consist of multiple levels. If a document output with errors is processed repeatedly, the output process will be aborted. By default, a new output will then be restarted in the step in which the output was aborted. This way, the document is not transferred to downstream systems only to be looped back around to them again.
For output, the output format and the output system are mutually independent, with the identification of the target system proceeding independently of output format generation. The output system designates the target system for data transfer (e.g., file system, mail server, or database). The output system must therefore always be configured.
Configuring an output format, on the other hand, causes an additional file attachment to be generated in just that format. If supported by the format, other existing file attachments can also be embedded. For example, if an XML file is to be output to the file system, the process would be divided in two stages, the first being definition of the output format "XML" and the second being transfer to the output system "File." In other words, the first step generates a new XML file from the index data, and the second step physically writes the file to the file system.
It is not always useful for an index file to be both output and generated. With most programming interfaces, the index data is not transferred in file form. As an alternative, an index file from the input data can be transferred unchanged. In terms of configuration syntax, multiple output formats and output systems can be defined in any order. The main aspect to note here is that the format must be generated before the actual transfer.
Sometimes, rather than outputting the documents synchronously step by step, output workers process documents by adding an asynchronous interval for fulfilling a condition. An asynchronous interval is especially used for output systems that receive data from xSuite Interface but do not process it synchronously and do not immediately provide a status feedback.
In order to receive the final processing status of this data, the program must periodically query the output system as to whether the processing there has taken place and what the result was. Only then can the processing also be continued in xSuite Interface with the next output step or with the final backup step. Asynchronous output steps are treated and configured in xSuite Interface like a separate pseudo-output system.
Output via a web service is yet another special case. It is important to note that an output step itself represents a web service rather than a normal output worker step. This case of processing is asynchronous because, before processing can continue in xSuite Interface, the Output Web Service must wait for an external process to request and retrieve the data that is waiting for output.