how to read the parameter and parameter value from the text file

Hii everyone.. I want to know how to read the text file which contains a variable and its value and then can be used somewhere in the program of .m matlab file... for eg.. example.txt is the text file... whcich contains like this... 110.01 voltage 230 volts 110.02 Current 1.094 amps

now I want to read 110.01 & 110.02 as variable and 230 & 1.094 as their values respectively... but these values present in somewhere in the text... how to identify these parameters and their values... please anyone suggest the command which I have to use for the above said problem

regards Madhu

Answer 1

[Stephen23]

The simplest solution is to read all of the data as a string, and split this into a cell array using regexp:

S = fileread('temp.txt');
C = regexpi(S,'^(\d+\.\d+) (.{20}) (.+?)$','tokens','lineanchors');
C = strtrim(vertcat(C{:}));

That's it! The value 20 is the width of the middle column. The variable C has three columns: the first has strings of the first values, the second has descriptions, and the third is the values. For example we can see the third row like this:

>> C(3,:)
ans = 
  '112.03'    'INU MAX POWER'    '9000 kW'

If you want to convert the values to numeric, then you can easily do this in one step using my FEX submission sinum, which converts string like '9000 kW' to the numeric value of 9000000 plus the unit string 'W'. You can use it simply like this:

[val,spl,sgf] = cellfun(@sinum,C(:,3),'UniformOutput',false);

Again looking at the third row, we already have its value and unit:

>> val{3}
ans =
     9000000
>>  spl{3}{2}
ans = 
    'W'

This code works correctly on the following data file:

[madhu T S]

Hello Stephen Cobeldick .. the code which you gave is really working well.. but I'm using Matlab2009a, so sinum code is not working..

regards madhu

[Stephen23]

As I wrote in my answer, sinum is not inbuilt code, and to use it you will need to download it from here:

http://www.mathworks.com/matlabcentral/fileexchange/33174-prefixed-string-conversion--si-or-binary-

I know it works on 2009a, because I wrote it and tested it.

[madhu T S]

ohh okiee Stephen.. thank you can v make selected text as a seperate file??? for example.. 150 series text (150.01 to 150.23)

[Stephen23]

Sorry, I don't understand the question. Which "separate file" ? None of the ID numbers in the data that you gave are 150, so where do these numbers come from?

[madhu T S]

Hi.. Stephen.. sorry.. let me make it clear.... the txt file contains several ID numbers such as 110.01, 110.02, 120.01, 121.01... 150.01.. 150.05... 160.05 and so on... each ID hold a value.. now if I want to select/read only 150. series ID's and their values and save as .mat file... how can I save?? Hope its clear.. let me know if u dont understand

regards madhu

[Stephen23]

You can easily select a subset of your data that matches any conditions that you want. If you want to select all values starting with '150', then using strncmp gives the indices of those rows:

idx = strncmp(C(:,1),'150',3)

You can use this to extract just those rows:

C(idx,:)

and

val(idx)

etc.

You could even concatenate those numeric values into a single numeric array like this:

[val{idx}]

and then do whatever further processing you need. You can save to .MAT file using save.

Answer 2

[Guillaume]

You're a bit vague about the structure of your text file. Is it all on one line? Are there any other numbers in your file? Does the text matter or just the number? What is the format of the number (are numbers like 1e4, NaN, -1246 possible?

Possibly, this will work for you:

numbers = str2double(regexp(fileread('example.txt'), '(?<=\s+|^)[+-]?\d*\.?\d+(?=\s+|$)', 'match'));
variables = numbers(1:2:end);
values = numbers(2:2:end);

Note: the regular expression will find any number enclosed by whitespace (or beginning or end of string) that conform to the following pattern: an optional + or - sign, followed by 0 or more digits, followed by an optional decimal separator, followed by 1 or more digits. That is the following are considered numbers:

-1.5
+5.6
-1
-.3
.6
1.78
8

The following are not

1.
1e6
NaN
Inf

[madhu T S]

thank you for your comments Guillaume, I have so much of data in txt file in that I have to Identify the tag INVERTER DATA.. and it will be in the format as shown below.. right after the tag the parameter and its values starts..

   INVERTER DATA:
112.01 INU NOM VOLTAGE      3100 V
112.02 INU MAX CURRENT      1650 A
112.03 INU MAX POWER        9000 kW
112.04 NOM DC VOLTAGE       4840 V
112.05 INU CONFIGURATION    SINGLE INU
112.06 LSU SUPERVISION      OFF
112.08 HARDPARALLEL ENA     DISABLE
112.09 PROT SCALING MODE    AUTOMATIC
112.10 DC Link Capacity     0 uF
112.11 Series Inductance    0 mH
112.12 Stray Capacitance    ?????? (Real)
112.16 LOW FREQ MODE ENA    DISABLE
112.17 OV FIRE THRGH ENA    DISABLE
112.23 OPT MODULATOR ENA    ENABLE
112.25 EPLD RE-TRIG ENA     ENABLE
112.26 ENA ASYM CURR DET    ENABLE

I have to Identify the code (112.01 to 112.26) and its corresponding value(like 3100, 1650,9000 and so on) respectively.. the middle data is not necessary and the number of white space is not constant for all the lines... how can solve this problem??

[per isakson]

How is the end of INVERTER DATA: indicated? By "112.26" in the first column?

[Guillaume]

To add to per's question:

• Is the number of entries fixed?
• Is the text in the middle fixed for a given code
• What kind of whitespace is responsible for the alignment of the value column? a variable number of blank spaces or a tab character?
• Is the format of the value fixed for a given code
• When the value is a number, should the unit be discared?
• When the value is a number, is it always integer?
• When it's a string what should be stored?

Could you show what you'd want as an output for your example?

[madhu T S]

hello Guilaume.. In the txt file there are num of blocks.. one among them is INVERTER DATA:

The num of entries, the text in the middle, format is fixed.. the value only changes for diff txt files and it may be floating number also(decimals).. when it is a string.. it should be stored as it is displaying..

Answer 3

[Guillaume]

The following will find the INVERTER DATA line and then parse each line until it does not match the following format: two number separated by a dot followed by an arbitrary number of alphabetic characters or spaces followed by at least two spaces followed by an arbitrary number of characters:

function parsed_data = decode_file(filename)
   validateattributes(filename, {'char'}, {'row', 'nonempty'});
   parsed_data = {};
   fid = fopen(filename, 'rt');
   tline = fgetl(fid);
   while ischar(tline) && isempty(strfind(tline, 'INVERTER DATA'))
      tline = fgetl(fid);
   end
   if ~ischar(tline)
      error('end of file reach prematurely'); 
   end
   pattern = '^(\d+\.\d+)[A-Za-z ]+ {2,}(.*)$';
   tline = fgetl(fid);
   while ischar(tline)
      parsed_line = regexp(tline, pattern, 'tokens', 'once');
      if isempty(parsed_line) %not a code / value line
         break;  %stop processing
      end
      parsed_data = [parsed_data; parsed_line]; %#ok<AGROW>
      tline = fgetl(fid);
   end
   fclose(fid);
end

[madhu T S]

Hello Guilaume...

I tried to run the code given by u.. found error as shown below

??? Undefined function or method 'decode_file' for input arguments of type 'char'.

    Error in ==> example at 187
    parsed_data = decode_file(filename);

[Guillaume]

Typical reason for this is the function is not in your matlab path. It certainly has nothing to do with my code.