clone: Serialization of record linkage object.

In RecordLinkage: Record Linkage Functions for Linking and Deduplicating Data Sets

Serialization of record linkage object.

Description

Saving, loading and deep copying of record linkage objects for big data sets.

Usage

  clone(object, ...)
  saveRLObject(object, file, ...)
  loadRLObject(file)

Arguments

object	Object of class "RLBigData". The object to save.
file	The name of the file to save to or load from.
...	Optional arguments for possible additions, currently not used.

Details

The classes for big data sets make use of file-backed data structures from the ff package, therefore the load and save mechanism of R is not sufficient for persistent storage of these objects. Also, assignment via <- does not duplicate the ff data structures.

clone makes a deep copy of an object by duplicating the underlying files.

saveRLObject saves an object to zip file containing a dump of the R object as well as the associated ff files.

loadRLObject loads an object from a file saved by saveRLObject.

clone and saveRLObject are generic functions with methods for "RLBigData" and "RLResult".

If loadRLObject is called with inPlace = FALSE (the default), a working copy of the database is made in a temporary file and the original file left untouched. Calling with inPlace = TRUE sets the provided file as working copy and changes made to the database are persistent. This option is useful when working with large files in order to prevent disk usage overhead.

saveRLObject and loadRLObject require working zip / unzip programs.

Value

clone returns a deep copy of object.

loadRLObject returns the loaded object.

saveRLObject is used for its side effects.

Note

Objects loaded with inPlace = TRUE must be saved again after changes have been made to the object (e.g. calculation of weights).

Author(s)

Andreas Borg, Murat Sariyar

RecordLinkage documentation built on Nov. 10, 2022, 5:42 p.m.