Categories

Versions

You are viewing the RapidMiner Studio documentation for version 10.2 - Check here for latest version

Collect and Persist (Operator Toolbox)

Synopsis

This operator is used for collecting various objects created during the execution of a process. It will generally used in combination with the ''Recall from App'' operator.

Description

This is a variation to the Collect operator. It basically is a combination of Remember and Collect operator. It creates a new collection object when this operator gets executed the first time and automatically saves it in cache. Subsequent use of the same operator with the same object name will add more objects to the collection.

The collection result can then be recovered using a simple ''Recall from App'' operator. Uncheck the ''remove from app'' option on the "Recall from App" operator to ensure the cache is cleared between successive runs of the process.

Input

  • input (Data Table)

    Any object can be provided here. This object will be appended to the collection which is stored in the object store of the process under the name specified by the 'objectname' parameter.

Output

  • output (Collection)

    The collection saved in cache.

Parameters

  • objectname The name under which the input object is stored is specified through this parameter. This name should be the same name that can be used later by the Publish to App operator. Range:
  • naming_convention Objects are assigned with this additional string to ensure that they can be easily identified later on. Macros can be used for this string. Range:

Tutorial Processes

Collect and Persist model and performance outputs

This tutorial process demonstrates feature selection optimization using the Collect and Persist operator. It saves all models and performance vectors in two separate collections. Selecting the "remove from app" parameter on the Recall from App operator causes different effects on multiple runs of the same process.