8.5

Splitters initiate a recurring cycle that stops only when the original file has been completely processed.
When a given splitter creates a file, it hands it down to the task that follows, and all the tasks on the same
branch are performed until the output task. Then the splitter task creates yet another file that is again handed
down to the next task, and so forth until the cycle ends (when there is no more data in the original file).
If the process merges the split data with a document, the splitter must not alter the structure of the data file. In
other words, each split file must have the same structure as the original files, otherwise the PlanetPress
Design documents to which they will be sent will not be able to extract the data correctly and the merging
process will fail.
Warning
Splitters do not modify the metadata that is currently active within your process. This means that, if you
are intending to use metadata along with a process using splitters, you can either use the "Metadata
Sequencer" on page464 instead of a splitter, or (re)create the metadata after the splitter.
About Using Emulations with Splitters
When an emulation is used with a splitter action task, the job file is emulated, cut to pieces and de-emulated.
Most times, the emulation/de-emulation process is completely transparent. However, in some cases, there
may be minute differences.
When using the ASCII or Channel Skip emulation, if there are missing line feed characters (when lines end
with a single carriage return in ASCII, or when lines start with a No line feed channel in Channel Skip), the
output data will be different from the input data, but the change will not be significant.
Let us imagine that a splitter action task processes the following data file using the ASCII emulation:
Data line1 of page 1<cr><lf>
Data line2 of page 1<cr>
Last data line of page 1<cr><lf>
Data line1 of page 2<cr><lf>
...and so forth...
Once split, the first file generated by the action task would look like this: