Data sinks are generally used to catch the output of a model. There are three forms of data sink: text-based sinks (described here), tabular sinks and graph sinks. In the case of text-based sinks, messages sent to the sink are added to the end of data sink's contents file, together with an index showing the cycle number when the message was received. Initialisation results in the inclusion of a data stamp within the sink's text file.
There is no limit to the number of data sinks you can have in one model, and multiple data sinks will behave as independent devices.
All data sinks have three properties:
File [values: an arbitrary character string]
This property specifies the name of the output file in which data is stored. It is possible to capture data from different runs of the same model in different output files.
Location [values: Local/IO Directory; default: Local]
This property specifies the location of output files. Sink output may be stored either locally (with the definition of the sink), or in the standard I/O directory, from where it may be retrieved (e.g., for inclusion in a document).
Accumulate [values: No/Yes; default: No]
This property controls sink behaviour on successive runs of a model. If this property is set to "Yes", data from previous runs will be not be discarded when the model is reinitialised. Instead, new data will just be added on, either at the bottom of the file (in the case of text-based sinks), or in the table (in the case of tabular sinks).