3.12

TransformMD v1

../_images/TransformMD-v1_dlg.png

TransformMD dialog.

Summary

Scale and/or offset the coordinates of a MDWorkspace

Properties

Name Direction Type Default Description
InputWorkspace Input IMDWorkspace Mandatory Any input MDWorkspace.
Scaling Input dbl list 1 Scaling value multiplying each coordinate. Default 1. Either a single value or a list for each dimension.
Offset Input dbl list 0 Offset value to add to each coordinate. Default 0. Either a single value or a list for each dimension.
OutputWorkspace Output IMDWorkspace Mandatory Name of the output MDWorkspace.

Description

This algorithm applies a simple linear transformation to a MDWorkspace or MDHistoWorkspace. This could be used, for example, to scale the Energy dimension to different units.

Each coordinate is tranformed so that x'_d = (x_d * s_d) + o_d where:

  • d : index of the dimension, from 0 to the number of dimensions
  • s : value of the Scaling parameter
  • o : value of the Offset parameter.

You can specify either a single value for Scaling and Offset, in which case the same m_scaling or m_offset are applied to each dimension; or you can specify a list with one entry for each dimension.

Notes

The relationship between the workspace and the original MDWorkspace, for example when the MDHistoWorkspace is the result of BinMD v1, is lost. This means that you cannot re-bin a transformed MDHistoWorkspace.

No units are not modified by this algorithm.

Performance Notes

  • Performing the operation in-place (input=output) is always faster because the first step of the algorithm if NOT in-place is to clone the original workspace.
  • For MDHistoWorkspaces done in-place, TransformMD is very quick (no data is modified, just the coordinates).
  • For MDWorkspaces, every event’s coordinates gets modified, so this may take a while for large workspaces.
  • For file-backed MDWorkspaces, you will find much better performance if you perform the change in-place (input=output), because the data gets written out to disk twice otherwise.

Categories: Algorithms | MDAlgorithms\Transforms

Source

C++ source: TransformMD.cpp (last modified: 2018-02-27)

C++ header: TransformMD.h (last modified: 2018-02-14)