Publishing streams

Publishing streams is done directly from IBM® SPSS® Modeler using any of the standard export nodes: Database, Flat File, Statistics Export, Extension Export, Data Collection Export, SAS Export, Excel, and XML Export nodes. The type of export node determines the format of the results to be written each time the published stream is executed using the IBM SPSS Modeler Solution Publisher Runtime or external application. For example, if you want to write your results to a database each time the published stream is run, use a Database export node.

To publish a stream

  1. Open or build a stream in the normal manner and attach an export node at the end.
  2. On the Publish tab in the export node, specify a root name for the published files (that is, the file name to which the .pim, .par, and .xml extensions will be appended).
  3. Click Publish to publish the stream, or select Publish the stream to automatically publish the stream each time the node is executed.

Published name. Specify the root name for the published image and parameter files.

  • The image file (*.pim) provides all of the information needed for the Runtime to execute the published stream exactly as it was at the time of export. If you are confident that you will not need to change any of the settings for the stream (such as the input data source or the output data file), you can deploy the image file only.
  • The parameter file (*.par) contains configurable information about data sources, output files, and execution options. If you want to be able to control the input or output of the stream without republishing the stream, you will need the parameter file as well as the image file.
  • The metadata file (*.xml) describes the inputs and outputs of the image and their data models. It is designed for use by applications which embed the runtime library and which need to know the structure of the input and output data.
    Note: This file is only produced if you select the Publish metadata option.

Publish parameters. If required, you can include stream parameters in the *.par file. You can change these stream parameter values when you execute the image either by editing the *.par file or through the runtime API.

This option enables the Parameters button. The Publish Parameters dialog box is displayed when you click the button.

Choose the parameters you want to include in the published image by selecting the relevant option in the Publish column.

On stream execution. Specifies whether the stream is automatically published when the node is executed.

  • Export data. Executes the export node in the standard manner, without publishing the stream. (Basically, the node executes in IBM SPSS Modeler the same way it would if IBM SPSS Modeler Solution Publisher were not available.) If you select this option, the stream will not be published unless you do so explicitly by clicking Publish in the export node dialog box. Alternatively, you can publish the current stream using the Publish tool on the toolbar or by using a script.
  • Publish the stream. Publishes the stream for deployment using IBM SPSS Modeler Solution Publisher. Select this option if you want to automatically publish the stream every time it is executed.
Note:
  • If you plan to run the published stream with new or updated data, it is important to note that the order of fields in the input file must be the same as the order of fields in the source node input file specified in the published stream.
  • When publishing to external applications, consider filtering extraneous fields or renaming fields to conform with input requirements. Both can be accomplished using a Filter node prior to the export node.
Parent topic: Extension Export node