Scanning a text file for bits and pieces of information

Hi.

I'm new to matlab and I'm a bit stuck in where to begin with coding a program to open a text file, read specific strings and extract information from specific strings. Here's my general outline and an example of the text format I'm trying to read.

Example format:(wanted information has asterisks on either side of it)

     <solution solution_id="telemetry" name="SITE_FRAME" add_date="2004-01-26T12:04:23Z" index1="1">
              <reference_frame name="SITE_FRAME" index1=" *0* "/>
              <offset x=" *-0.0* " y=" *0.0* " z=" *0.0* "/>
              <orientation s="1.0" v1="0.0" v2="0.0" v3="0.0"/>
          </solution>
          <alias>
              <old index1="0" index2="0" index3="1" index4="59"/>
              <new index1="1"/>
          </alias>
      <solution solution_id="telemetry" name="SITE_FRAME" add_date="2004-01-26T12:04:25Z" index1="2">
              <reference_frame name="SITE_FRAME" index1=" *1* "/>
              <offset x=" *-0.0* " y=" *0.0* " z=" *0.0* "/>
              <orientation s="1.0" v1="0.0" v2="0.0" v3="0.0"/>
          </solution>
          <alias>
              <old index1="1" index2="0" index3="0" index4="13"/>
              <new index1="2"/>

    Outline:
    Open file 'mer_site' (that's the file that contains this format of information)
        Look for instances of * information
            ie the parts that say 'name=SITE_FRAME" index1="*"/>'
                 export the * information into a vector (total size [1x158])
        Look for instances of * information
            ie the parts that say '<offset x="*" y="*" z="*"/>'
                 export the * information into a vector (total size [3x158])
    Close File

   So far, all I know is that I need
    fid=fopen('mer_site')
     something with textscan
     something about putting the results from textscan into a cell 
    closing the file

I'm not sure which arguments I need for textscan because the information is mixed.

I'd be super grateful if anyone could help with this!!!

[Walter Roberson]

Would you be able to use an xml parser on this?

[per isakson]

See http://se.mathworks.com/matlabcentral/answers/225199-read-text-file-and-identify-certain-values-line-by-line-or-character-by-character#answer_183899

[Elena H.]

I tried using it and maybe it's my beginner level usage of matlab, but nothing seems to be in the structure?

[per isakson]

You show an excerpt of a file and we guess that it is an xml-file. We are probably mistaken.

[Elena H.]

I originally got it as an svf file and I just saved it as a .txt file so I could upload it

[Walter Roberson]

That file does not appear to be in Simple File Verification (SFV) format. Perhaps it has something to do with State Fusion Vector ?

[Elena H.]

I downloaded the file from

http://pds-geosciences.wustl.edu/mer/mer1-m-eng-6-rmc-ops-v1/mer1rm_0xxx/data/

The file is mer1_master

[per isakson]

I found an "issue" with mer1_master.

<solution solution_id="telemetry" name="SITE_FRAME" add_date="2013-12-03T17:01:12Z" index1="182">
        <reference_frame name="SITE_FRAME" index1="181"/>
        ....
<solution add_date="2014-02-16T19:02:11Z" index1="183" name="SITE_FRAME" solution_id="telemetry">
        <reference_frame index1="182" name="SITE_FRAME"/>
        ....

The order of the "items" changes after index, 183. I guess that is not significant with an xml-file.

[Walter Roberson]

The directory above says that the files are indeed XML files.

Answer 1

[per isakson]

I'm surprised that xml2struct by Wouter Falkena failed with your file. Did it throw any error or warning message?

Instead of trying myself I did an exercise with regular expression.

>> out = cssm('c:\m\cssm\mer1_master.txt')
out = 
1x191 struct array with fields:
    index1
    x
    y
    z
>> out(5)
ans = 
    index1: 4
         x: 12.3513
         y: 4.1437
         z: -0.8949
>> out(185)
ans = 
    index1: 184
         x: -438.7025
         y: -0.5040
         z: -10.6850
>>

where

function  out = cssm( filespec )
    str = fileread( filespec );
    xpr = '(?<=<solution).+?(?=</solution>)';
    cac = regexp( str, xpr, 'match' );
    %
    out = struct( 'index1', num2cell(nan(1,length(cac)))...
                , 'x',[], 'y',[], 'z',[]                ); 
    %
    xpr = cat( 2                                    ...
                , '<reference_frame'                ...
                ,       '.*'                        ...  
                ,       ' index1="(?<index1>\d+)"'  ...
                ,       '.*'                        ...  
                , '/>'                              ...
                , '\s*'                             ...
                , '<offset'                         ...
                ,       ' x="(?<x>[\-\d\.]+)"'      ...
                ,       ' y="(?<y>[\-\d\.]+)"'      ...
                ,       ' z="(?<z>[\-\d\.]+)"'      ...
                , '/>'                              );
    %
    for jj = 1 : length( cac )
        sas = regexp( cac{jj}, xpr, 'names' );
        out(jj).index1 = str2double( sas.index1 );   
        out(jj).x      = str2double( sas.x      );   
        out(jj).y      = str2double( sas.y      );   
        out(jj).z      = str2double( sas.z      );   
    end
end

Caveat: This function is based on backward engineering of one single file and tested with the same file. It may fail sometime in the future with some other file.

[Elena H.]

This worked!!! Thank you so much! I'm super psyched and will give you credit if I ever publish the results of my data or reference this program. Thanks! I tried using something to process xml files and got an extremely confusing structure that I was trying to sift through, bit by bit. This helped so much!