Dataset with wrong x-Axis

SupportCommunity FPScript Dataset with wrong x-Axis

Ansicht von 2 Beiträgen - 1 bis 2 (von insgesamt 2)
  • Autor
    Beiträge
  • #28485
    Leon DLeon D
    Teilnehmer

    Hello,

    I have a signal that has a wrong x-axis. The time intervall between two maximas should be the same, but it isn`t. 

    To correct this, i tried to use given fpscript formulas and manipulated them. My Idea is to extract maxima, delete x-values in between two maxima and replace the x-values at the maxima with the x-value at the maxima before plus a fixed time intervall. (x_before + 50000s)

    The codes are modified fpscript codes from event isolation. I added the_X variable and swapped some things.

    This code is supposed to delete all x-values between to maxima and replace the x-value at the maxima with the x-value at the maxima before + 50000s. The deleting of x-values in between works fine, but i have no idea how to replace the maximum values with an ascending value.

    Dim _EventResult, _Sig, _Max, _Y, _Event, _X
    _Y = FloatingPoint64 SourceData.Y
    _X = FloatingPoint64 SourceData.X
    _Sig = '2_Temp_Thermistor'
    _Max = Maximum(_Sig)
    If DataStructure(_Max) >= STRUCT_DATASERIES Then
    _Max = Maximum(_Max)
    End
    _Event = Extrema(_Sig, 0.01000000000000000 * _Max, EVENT_POSITIVE, EVENT_INDEX)
    _X[IndexNot(_Event, NumberOfRows(_X))] = ?
    If DataStructure(SourceData) <= STRUCT_DATAMATRIX Then
    _EventResult = _Y
    Else
    _EventResult = Signal(SourceData.Y, _X, SourceData.Z)
    End
    End

    Interpolating afterwards works also fine.

    Thanks in advance!

     

     

    • Dieses Thema wurde geändert vor 1 Jahr, 7 Monaten von Leon DLeon D.
    • Dieses Thema wurde geändert vor 1 Jahr, 7 Monaten von Leon DLeon D.
    #28489
    Bernhard KantzBernhard Kantz
    Moderator

    The _Event variable contains a dataset with the indices of the local maximas. Thus _Event[0] marks the first, _Event[1] the second one. Thus

    _X[_Event[1]] = _X[_Event[0]] + 50000 s

    should do the trick.

     

Ansicht von 2 Beiträgen - 1 bis 2 (von insgesamt 2)
  • Du musst angemeldet sein, um auf dieses Thema antworten zu können.