ConjoinSpectra v1

../_images/ConjoinSpectra-v1_dlg.png

ConjoinSpectra dialog.

Properties

Name Direction Type Default Description
InputWorkspaces Input string Mandatory Comma seperated list of workspaces to use, group workspaces will automatically include all members.
OutputWorkspace Output Workspace Mandatory Name the workspace that will contain the result
WorkspaceIndex Input number 0 The workspace index of the spectra in each workspace to extract. Default: 0
LabelUsing Input string   The name of a log value used to label the resulting spectra. Default: The source workspace name
LabelValue Input string Mean How to derive the value from a time series property. Allowed values: [‘Mean’, ‘Median’, ‘Maximum’, ‘Minimum’, ‘First Value’]

Description

This algorithm allows a single spectrum to be extracted from a range of workspaces and placed into a single workspace for comparison and plotting. The LabelUsing property allows you to control what the end labels applied to each spectra will be. The default is to use the source workspace name, but you can specify the name of a log value to use as the label, e.g. Temp_Sample. the LabelValue property allows control of how a single value is extracted from time series logs.

ConjoinSpectra operation

Example case with 3 input workspaces. ConjoinSpectra operation

Usage

Example - Join workspace index 1 from three workspaces

ws = CreateSampleWorkspace("Histogram",BankPixelWidth=1)
ws2 = CreateSampleWorkspace("Histogram",BankPixelWidth=1)
ws3 = CreateSampleWorkspace("Histogram",BankPixelWidth=1)

wsOut = ConjoinSpectra("ws,ws2,ws3",WorkspaceIndex=1)
print ("Creates a workspace with %i spectra labelled as" % wsOut.getNumberHistograms())
print wsOut.getAxis(1).extractValues()

Output:

Creates a workspace with 3 spectra labelled as
['ws_1', 'ws2_1', 'ws3_1']

Example - Join spectra from three workspaces using a log value

ws = CreateSampleWorkspace("Histogram",BankPixelWidth=1)
ws2 = CreateSampleWorkspace("Histogram",BankPixelWidth=1)
ws3 = CreateSampleWorkspace("Histogram",BankPixelWidth=1)

#add some log values
wsIndex=0
for w in [ws,ws2,ws3]:
    wsIndex += 1
    for minute in range(60):
        time_string = "2010-01-01T00:%02i:00" % minute
        AddTimeSeriesLog(w, Name="myLog", Time=time_string, Value=(wsIndex*100)+minute)

wsOut = ConjoinSpectra("ws,ws2,ws3",WorkspaceIndex=0,LabelUsing="myLog",LabelValue="Mean")
print ("Creates a workspace with %i spectra labelled using the Mean of myLog." % wsOut.getNumberHistograms())
print wsOut.getAxis(1).extractValues()

otherLabelValueOptions = ["First Value","Median", "Maximum", "Minimum"]
for labelValueOption in otherLabelValueOptions:
    wsOut = ConjoinSpectra("ws,ws2,ws3",WorkspaceIndex=0,LabelUsing="myLog", LabelValue=labelValueOption)
    print ("Creates a workspace with %i spectra labelled using the %s of myLog." % (wsOut.getNumberHistograms(), labelValueOption))
    print wsOut.getAxis(1).extractValues()

Output:

Creates a workspace with 3 spectra labelled using the Mean of myLog.
['129.5', '229.5', '329.5']
Creates a workspace with 3 spectra labelled using the First Value of myLog.
['100.0', '200.0', '300.0']
Creates a workspace with 3 spectra labelled using the Median of myLog.
['129.5', '229.5', '329.5']
Creates a workspace with 3 spectra labelled using the Maximum of myLog.
['159.0', '259.0', '359.0']
Creates a workspace with 3 spectra labelled using the Minimum of myLog.
['100.0', '200.0', '300.0']

Categories: Algorithms | Transforms\Merging