12 Common patterns

This section outlines common development patterns encountered while developing operators in Tercen.

12.1 Output data

12.1.1 One table

Coming soon.

Example: Mean operator and lm operator.

12.1.2 List of tables

Coming soon.

Example: PCA operator.

12.1.3 Serialised R object

Coming soon.

Example: FlowSOM operator.

12.1.4 Serialised file

Coming soon.

Example: Barplot operator.

12.2 Output relations

12.2.1 Per Cell

Example: Mean operator, lm operator.

12.2.2 Per Row / Per Column

There are two typical cases for creating a relation per row or per column only:

  • the output is computed per observation. This is the case of clustering algorithms that assign each observation to a cluster ID.

  • the is computed per variable. For example, we want to scale our values and perform, this operation per variable or stratifying factor in our dataset.

Example: PCA operator.

12.2.3 No relation

This is useful when no relation needs to be made between computed values and existing data. For example, we can consider two cases where this is relevant:

  • the operator computes an end result that cannot be linked to individual input data points
  • the computation must be available to be linked to any data point (for example, when we want to generate a constant that can be used together with any other data)

Example: Constant operator.