When the build
command is executed, all of the data
processing commences. Any pre-processing necessary within the table
environment takes place first. Next, each of the layers begins executing.
Once the layers complete executing, the output of each layer is stacked into
the resulting data frame.
Once this process is complete, any post-processing necessary within the table
environment takes place, and the final output can be delivered. Metadata and
traceability information are kept within each of the layer environments,
which allows an investigation into the source of the resulting datapoints.
For example, numeric data from any summaries performed is maintained and
accessible within a layer using get_numeric_data
.
The `metadata` option of build will trigger the construction of traceability
metadata for the constructed data frame. Essentially, for every "result" that
Tplyr produces, Tplyr can also generate the steps necessary to obtain the
source data which produced that result from the input. For more information,
see vignette("metadata").