.. _Split.json#/:

Split
=====

| Splits the data stream into multiple different streams.
  Processes each split stream with different settings.
| *Warning:* The split step must be the last step in a list of successive steps.
  Specifying steps that follow the 'split' step is only possible using
  the 'steps' settings of the split step. For example, use
  ``steps=[predict,split] split.steps=[average,out]`` `.`


.. _Split.json#/inputs:

inputs
******


.. _Split.json#/inputs/type:

type
++++

:type: ``string``

:doc: Case-insensitive step type; must be 'split' `.`


.. _Split.json#/inputs/replaceparms:

replaceparms
++++++++++++

:type: ``array of strings``

:default: ``[]``

:doc: A list of key names that are different for each of the next steps. The keys that are specified in this list should themselves be given as a list of parameters. E.g., for ``split.replaceparms=[average.timestep, msout.name]``, ``average.timestep`` and ``msout.name`` should now be specified as arrays, e.g. ``average.timestep=[1, 4]``. The keywords that are specified by split.replaceparms determine the number of streams that the input is split into. In the above example, there will be two streams because ``average.timestep`` lists two values. All other specified keywords should also list two values `.`


.. _Split.json#/inputs/steps:

steps
+++++

:type: ``array of strings``

:default: ``[]``

:doc: List steps that follow the 'split' step, e.g. ``[average, msout]``. Split creates multiple pipelines with these steps, using the replaceparms setting `.`