creating a strucure from a string from my Tektronix MDO3034 oscilloscope

Hi

I get this string from my oscilloscope

:WFMPRE:BYT_NR 2;BIT_NR 16;ENCDG ASCII;BN_FMT RI;BYT_OR MSB;WFID "Ch1, DC coupling, 50.00mV/div, 20.00us/div, 10000 points, Sample mode";NR_PT 10000;PT_FMT Y;XUNIT "s";XINCR 20.0000E-9;XZERO -80.8000E-6;PT_OFF 0;YUNIT "V";YMULT 7.8125E-6;YOFF -14.7200E+3;YZERO 0.0E+0;VSCALE 50.0000E-3;HSCALE 20.0000E-6;VPOS -2.3000;VOFFSET 0.0E+0;HDELAY 19.2000E-6;DOMAIN TIME;WFMTYPE ANALOG;CENTERFREQUENCY 0.0E+0;SPAN 0.0E+0;REFLEVEL 0.0E+0

and I want to create a structure wfmpre that splits it up like this.

wfmpre.WFMPRE_BYT_NR='2'
wfmpre.BYT_NR='2'
wfmpre.ENCDG='ASCII'
wfmpre.BN_FMT='RI'
wfmpre.BYT_OR='MSB'
wfmpre.WFID='Ch1, DC coupling, 50.00mV/div, 20.00us/div, 10000 points, Sample mode'

You get the idea... But how can I do it?

Answer 1

[Stephen23]

Using textscan makes this task easy, and also allows us to use the '%q' format specifier so that double-quoted strings are kept together (even if they contain delimiter characters).

S = ':WFMPRE:BYT_NR 2;BIT_NR 16;ENCDG ASCII;BN_FMT RI;BYT_OR MSB;WFID "Ch1; DC coupling, 50.00mV/div, 20.00us/div, 10000 points, Sample mode";NR_PT 10000;PT_FMT Y;XUNIT "s";XINCR 20.0000E-9;XZERO -80.8000E-6;PT_OFF 0;YUNIT "V";YMULT 7.8125E-6;YOFF -14.7200E+3;YZERO 0.0E+0;VSCALE 50.0000E-3;HSCALE 20.0000E-6;VPOS -2.3000;VOFFSET 0.0E+0;HDELAY 19.2000E-6;DOMAIN TIME;WFMTYPE ANALOG;CENTERFREQUENCY 0.0E+0;SPAN 0.0E+0;REFLEVEL 0.0E+0';
C = textscan(S,'%q','Delimiter',{';',' '});
C = reshape(C{1},2,[]);
C(1,:) = regexprep(C(1,:),{'^[^A-Z]+','\W'},{'','_'},'ignorecase');
out = struct(C{:})

creates this structure:

out = 
      WFMPRE_BYT_NR: '2'
             BIT_NR: '16'
              ENCDG: 'ASCII'
             BN_FMT: 'RI'
             BYT_OR: 'MSB'
               WFID: 'Ch1; DC coupling, 50.00mV/div, 20.00us/div, 10000 points, Sample mode'
              NR_PT: '10000'
             PT_FMT: 'Y'
              XUNIT: 's'
              XINCR: '20.0000E-9'
              XZERO: '-80.8000E-6'
             PT_OFF: '0'
              YUNIT: 'V'
              YMULT: '7.8125E-6'
               YOFF: '-14.7200E+3'
              YZERO: '0.0E+0'
             VSCALE: '50.0000E-3'
             HSCALE: '20.0000E-6'
               VPOS: '-2.3000'
            VOFFSET: '0.0E+0'
             HDELAY: '19.2000E-6'
             DOMAIN: 'TIME'
            WFMTYPE: 'ANALOG'
    CENTERFREQUENCY: '0.0E+0'
               SPAN: '0.0E+0'
           REFLEVEL: '0.0E+0'

To convert the strings with number values to numeric, then replace the last line with this:

num = cellfun(@str2double,C(2,:));
idx = ~isnan(num)|strcmpi('NaN',C(2,:));
C(2,idx) = num2cell(num(idx));
out = struct(C{:})

which creates this output:

out = 
    WFMPRE_BYT_NR: 2
           BIT_NR: 16
            ENCDG: 'ASCII'
           BN_FMT: 'RI'
           BYT_OR: 'MSB'
           ... etc

[Micke Malmström]

Thanks allot! it works fine!

[Stephen23]

My pleasure. You can also accept the answer that best resolves your question. This is a small sign of appreciation for us.

[Micke Malmström]

Is this updated code with "textscan" is faster than the previous one with only regexp? Or should they be the same?

[Stephen23]

I did not test the timing, the advantage of textscan is that it is much more robust for handling double-quoted strings and is also much easier to understand. It is possible that it is faster.