八度：用 'textscan' 替换 'strread' - 抛出“错误：返回列表中未定义的元素号 4

Question

I am trying to run.m script in Octave. It runs well in matlab. a snippet is as follows

fid=fopen(log_file);
number = 1;
while 1
    tline = fgetl(fid);
    if ~ischar(tline), break, end
    
    [time1,time2,time3, string1, string2, line_tag, data] = strread(tline,'%d%d%d%s%s%s%[^\n]','delimiter',': ');

Octave throws the following error

log_file = ods_rlog.log
warning: strread is obsolete; use textscan instead
error: strread: %q, %c, %[] or bit width format specifiers are not supported yet.
error: called from
    strread at line 349 column 7
    plot_ods_rlog_s at line 35 column 59

As per suggestion I replaced strread with textscan it gives me the following error.

log_file = ods_rlog.log
error: element number 4 undefined in return list
error: called from
    plot_ods_rlog_s at line 35 column 59
>>

The file it reads from contains the following data

02:30:18 : PROODS: OBSTACLE: CFUNS 25 27 22 25
02:30:18 : PROODS: OBSTACLE: START_END_ATOM 12 14
02:30:18 : PROODS: OBSTACLE: ATOMS 12 19 17 1 6058 R 4545031 3682089 4550870 3682088 4550670 3684688 4545231 3684689 3360 0
02:30:18 : PROODS: OBSTACLE: ATOMS 13 20 24 1 6058 R 4545231 3684689 4550670 3684688 4550669 3780977 4545230 3780978 3360 0
02:30:18 : PROODS: OBSTACLE: ATOMS 14 26 27 0 6058 R 4545230 3780978 4550669 3780977 4550670 3793976 4545231 3793977 3360 0

Any suggestions as to what could be the problem.

Thanks !

Answer 1

You are using textscan wrong. The syntax (according to the documentation) is:

 -- C = textscan (FID, FORMAT)
 -- C = textscan (FID, FORMAT, REPEAT)
 -- C = textscan (FID, FORMAT, PARAM, VALUE, ...)
 -- C = textscan (FID, FORMAT, REPEAT, PARAM, VALUE, ...)
 -- C = textscan (STR, ...)
 -- [C, POSITION, ERRMSG] = textscan (...)

Ie it expects either a single output, or three. You are passing seven.

Capture what you require in a single output first, and split it later if necessary.

octave:1> datafile = fopen('data.colsv');
octave:2> C = textscan( datafile,'%d:%d:%d:%s:%s:%s%[^\n]');
octave:3> [time1,time2,time3, string1, string2, line_tag, data] = C{:}
time1 = 2
time2 = 30
time3 = 18
string1 =
{
  [1,1] = PROODS:
}

string2 =
{
  [1,1] = OBSTACLE:
}

line_tag =
{
  [1,1] = CFUNS
}

data =
{
  [1,1] = 25 27 22 25
}
    
    string2 =
    {
      [1,1] = OBSTACLE:
    }

    line_tag =
    {
      [1,1] = CFUNS
    }

    data =
    {
      [1,1] = 25 27 22 25
    }

However, textscan is a horrible command, and you should only use it because matlab doesn't have a better alternative (there are some csv2cell functions in matlabexchange though).

With textscan, you capture these line by line, and you'll have to loop.

A far better way is to use the csv2cell function from the io package. Eg, using your example above:

octave:4> pkg load io
octave:5> csv2cell ( 'data.colsv', ':' )
ans =
{
  [1,1] =  2
  [2,1] =  2
  [3,1] =  2
  [4,1] =  2
  [5,1] =  2
  [1,2] =  30
  [2,2] =  30
  [3,2] =  30
  [4,2] =  30
  [5,2] =  30
  [1,3] = 18 
  [2,3] = 18 
  [3,3] = 18 
  [4,3] = 18 
  [5,3] = 18 
  [1,4] =  PROODS
  [2,4] =  PROODS
  [3,4] =  PROODS
  [4,4] =  PROODS
  [5,4] =  PROODS
  [1,5] =  OBSTACLE
  [2,5] =  OBSTACLE
  [3,5] =  OBSTACLE
  [4,5] =  OBSTACLE
  [5,5] =  OBSTACLE
  [1,6] =  CFUNS 25 27 22 25
  [2,6] =  START_END_ATOM 12 14
  [3,6] =  ATOMS 12 19 17 1 6058 R 4545031 3682089 4550870 3682088 4550670 3684688 4545231 3684689 3360 0
  [4,6] =  ATOMS 13 20 24 1 6058 R 4545231 3684689 4550670 3684688 4550669 3780977 4545230 3780978 3360 0
  [5,6] =  ATOMS 14 26 27 0 6058 R 4545230 3780978 4550669 3780977 4550670 3793976 4545231 3793977 3360 0
}

You may have to process the final column a bit further with strsplit , or some combination of csvconcat / csvexplode but it's already easier to do so from here.