八度高级textread用法，bash

Question

I have following text file:

079082084072079032084069067072000000000,0 082078032049050032067072065082071069000,1 076065066032065083083084000000000000000,0 082078032049050072082000000000000000000,1 082078032049050072082000000000000000000,1 082078032049050072082000000000000000000,1 070083087032073073032080068000000000000,0 080067065032049050032072082000000000000,0 082078032056072082000000000000000000000,1 070083087032073073073000000000000000000,0 082078032087069069075069078068000000000,1 082078032049050072082000000000000000000,1 077065073078084032077069067072032073073,0 082078032049050072082000000000000000000,1 080067065032049050032072082000000000000,0 082078032049050072082000000000000000000,1

I need too matrices: X size 16x13 Y size 16x1

I want to separate each row of the file into 13 values, example: 079 082 084 072 079 032 084 069 067 072 000 000 000

Is it possible to import it into octave using textread function?

If no, can it be done using Linux bash command?

Answer 1

Yes, you can do this with textscan (see bottom if you really want to use textread :

octave> txt = "079082084072079032084069067072000000000,0\n082078032049050032067072065082071069000,1";
octave> textscan (txt, repmat ("%3d", 1, 13))
ans = 
{
  [1,1] =

    79
    82

  [1,2] =

    82
    78

  [1,3] =

    84
    32

  [1,4] =

    72
    49
[...]

Note that you are reading them as numeric values, so you do not get the preceding zeros. If you want them, you can either read them as string by using "%3s" in the format (extra trouble to handle and reduced performance since you will then be handling cell arrays).

Since you are reading from a file:

[fid, msg] = fopen ("data.txt", "r");
if (fid)
  error ("failed to fopen 'data.txt': %s", msg);
endif 
data = textscan (fid, repmat ("%3d", 1, 13));
fclose (fid);

If you really want to use textread :

octave> [d1, d2, d3, d4, d5, d6, d7, d8, d9, d10, d11, d12, d13] = textread ("data.txt", repmat ("%3d", 1, 13))
d1 =

  79
  82
  76
[...]    
d2 =

  82
  78
  65
[...]

or:

octave> data = cell (1, 13);
octave> [data{:}] = textread ("data.txt", repmat ("%3d", 1, 13))
data = 
{
  [1,1] =

    79
    82
    76
[...]

  [1,2] =

    82
    78
    65
[...]

If you need to capture the value after the comma (not really part of your original question), you can use:

octave> textscan (txt, [repmat("%3d", 1, 13) ",%1d"])
ans = 
{
  [1,1] =

    79
    82

  [1,2] =

    82
    78

  [1,3] =

    84
    32

[...]

  [1,14] =

0
1

}

Answer 2

You can do this pretty easily by reading three characters at a time using read in the shell:

while IFS="${IFS}," read -rn3 val tail; do
    [[ $tail ]] && echo || printf '%s ' "$val"
done < file

This implementation assumes that if we encounter a value after the comma, we should go to the next line.