How to deal with a nestled struct

Hi,

I have a xlm I wish to study. I used this code:

clc; clear all; close all

sampleXMLfile = 'A_AXIS_1.xml';
xDoc=parseXML(sampleXMLfile);

I need to manage data starting from line 87:

Browsing the struct i can find the data here

xDoc(2).Children(10).Children(2).Children(12).Attributes

Here i can find rec time, f1, f2 and f3 as Attributes.

How can i transfer them in a matrix so i can manipulate them?

[Walter Roberson]

Could you post a sample file for us to test with?

[andrea vironda]

Here the zip file

Basically the first part is useless to me, it is simply the setup environment.

t is in [s], f1 in [A], f2 in [°/min], f3 in [°]. It's an electrical test and i wish to plot the data

[Walter Roberson]

There are thousands of broken records in the file. For example,

  <rec time="5.842000" f1="-0.229889" f2="0.000671" f3="359.989757"/>
  <rec time="5.844000" f1="-0.232361" f3="359.989765"/>
  <rec time="5.846000" f2="-0.001341" f3="359.989749"/>
  <rec time="5.848000" f2="0.000447" f3="359.989754"/>
  <rec time="5.850000" f2="0.000000"/>

Notice the second line is missing f2, the third and fourth are missing f1, the fifth is missing f1 and f3.

What would you like to have happen in those cases? There are even some records that are missing time information.

[andrea vironda]

In those cases, i can accept an average between i-1 and i+1

[Walter Roberson]

In the subset I show, there are three records in a row with no f1, so taking mean of two adjacent records would not be able to fill the gap.

There are places that have over 75 missing f1 in a row.

When a defective record is encountered, can the entire record be discarded, and everything calculated based on only the full records? Or should as much be recovered as possible (a slower more complicated process)?

[andrea vironda]

I think discard the record is the best solution. In a second step i can think how to deal with missing parameters

Answer 1

[Walter Roberson]

sampleXMLfile = 'A_AXIS_1.xml';

S = fileread(sampleXMLfile);

just_numeric_cell = regexp(S, 'time="-?([\d.]+)" f1="-?([\d.]+)" f2="-?([\d.]+)" f3="-?([\d.]+)"', 'tokens');
just_numeric_array = vertcat(just_numeric_cell{:});
Attributes = str2double(just_numeric_array);

This finds about 6700 full records, with there being roughly another 2200 partial records in the file.

The largest time gap is about 0.07 seconds.

The order of columns is time, f1, f2, f3.