How to find a particular string in a text file

Hi, I have this text file ( Sample: SAM-3_Round-1_C-16_Ref_Spot-1 Psi= 43.309, Delta=105.412, No Soution

Sample: SAM-3_Round-1_C-16_Ref_Spot-2 Psi= 43.284, Delta=105.465, No Soution

Sample: SAM-3_Round-1_C-8_Spot-1 Psi= 43.266, Delta=107.861, No Soution

Sample: SAM-3_Round-1_C-8_Spot-2 Psi= 43.287, Delta=107.872, No Soution

Sample: SAM-3_Round-1_C-10_Spot-1 Psi= 43.269, Delta=106.890, No Soution

Sample: SAM-3_Round-1_C-10_Spot-2 Psi= 43.269, Delta=106.849, No Soution

Sample: SAM-3_Round-1_C-12_Spot-1 Psi= 43.267, Delta=106.872, No Soution

Sample: SAM-3_Round-1_C-12_Spot-2 Psi= 43.278, Delta=106.888, No Soution)

I want to search for the word 'C-8'(say), then I have to store the corresponding values of Psi and Delta (only the numeric values except special character). I tried "regexp" function for this but it didn't work. I am kind of beginner in MATLAB. It will be very helpful if someone can help me with this. Thank you.

[Stephen23]

Should "Soution" really be "Solution"?

Answer 1

[Stephen23]

It wold likely be much faster and simpler to read in the data normally, and then perform the search and matching inside of MATLAB, rather than trying to perform this on the file (or some string) and convert it afterwards.

Try using textscan, which is intended for this kind of filereading. In this example I named the file 'temp.txt', and also attached it below:

>> fid = fopen('temp.txt','rt');
>> C = textscan(fid,'%*s%s%s%f%s%f%[^\n]', 'Delimiter',',= ', 'MultipleDelimsAsOne',true);
>> fclose(fid);

You will find the all of the data in C:

>> C{1}
ans = 
  'SAM-3_Round-1_C-16_Ref_Spot-1'
  'SAM-3_Round-1_C-16_Ref_Spot-2'
  'SAM-3_Round-1_C-8_Spot-1'
  'SAM-3_Round-1_C-8_Spot-2'
  'SAM-3_Round-1_C-10_Spot-1'
  'SAM-3_Round-1_C-10_Spot-2'
  'SAM-3_Round-1_C-12_Spot-1'
  'SAM-3_Round-1_C-12_Spot-2'

This means you can quickly search for any substring (e.g. 'C-8') in this cell of strings, and then obtain the corresponding values from the other arrays:

>> idx = ~cellfun('isempty',strfind(C{1},'C-8'))
idx =
   0
   0
   1
   1
   0
   0
   0
   0

We can then use this index to extract all of the corresponding values of Psi and Delta (i.e. those corresponding to 'C-8'):

>> Psi = C{3}(idx)
Psi =
     43.266
     43.287
>> Delta = C{5}(idx)
Delta =
     107.86
     107.87

Note that you can easily combine different index requirements too, here we match any of idx and those with 'Spot-1':

>> idy = idx & ~cellfun('isempty',strfind(C{1},'Spot-1'))
idy =
   0
   0
   1
   0
   0
   0
   0
   0

Answer 2

[Soumya Bhattacharya]

Hi Stephen, Thank you so much for your help. When I run this code, Psi and Delta are not coming correctly. Psi =

   NaN

Delta =

   NaN
My actual file is attached bellow. It's scanning only first two lines. I am not sure how to scan the whole file. Can you please help me again ? Thank you again and have a nice day.
Soumya

[Stephen23]

Please use the comments for commenting on other answers or your own question. The Answers are supposed to be for actually answering the question. Note that the order of the answer can changes, so knowing what this "comment" applies to could be difficult in future.

[Stephen23]

The file uploaded has one group of data over two lines like this:

Sample: SAM-3_Round-1_C-16_Ref_Spot-1
Psi= 43.309, Delta=105.412, No Solution
Sample: SAM-3_Round-1_C-16_Ref_Spot-2
Psi= 43.284, Delta=105.465, No Solution
...

and also has many empty lines, which means that a basic textscan operation will not work. Here are two alternatives: one without the empty lines (using textscan), and with with the empty lines (using regexp).

1: If the empty lines are removed, then this will read the values from the file:

fid = fopen('temp.txt','rt');
C = textscan(fid,'%*s%s\n%*[^\n]');
fclose(fid);
fid = fopen('temp.txt','rt');
D = textscan(fid,'%s%f%s%f%[^\n]\n%*[^\n]', 'Delimiter',',=', 'HeaderLines',1);
fclose(fid);

And again we match any substring and get the required values:

>> idx = ~cellfun('isempty',strfind(C{1},'C-8'));
>> Psi = D{2}(idx)
Psi =
   43.2660
   43.2870
   43.2880
   43.3160
   43.3060
   43.3230
>> Delta = D{4}(idx)
Delta =
  107.8610
  107.8720
  107.8860
  107.9160
  107.8950
  107.9300

2: If the data file really must contain those (almost) emtpy lines, then this regexp will identify them out for further parsing:

str = fileread('temp.txt');
xpr = '(.+?)=(.+?),';
tkn = regexp(str,['\s*\S+:\s+(\S+)\s+',xpr,xpr,'([^\n]+)'],'tokens');
tkn = vertcat(tkn{:});

and the useful outputs are:

names = tkn(:,1);
Psi = cellfun(@(s)sscanf(s,'%f'),tkn(:,3));
Delta = = cellfun(@(s)sscanf(s,'%f'),tkn(:,5));

Answer 3

[Soumya Bhattacharya]

Hi Stephen, Thank you so much for your help again.I am using the 2nd method to read my data and it's working great. But, the problem is, I want to find Psi and Delta for C-8(say)(here it's giving me all the values for Psi and Delta), that's why I introduce the shorting line from method#1(idx = ~cellfun('isempty',strfind(tkn(:,1),'C-16'));). But I am not sure how can I apply this condition for Psi and Delta. Can you please kindly help me with this? I don't understand the function of xpr, it will be very kind of you if you can let me know about the function of that line. Thank you again for helping me. Have a nice day.

close all

clear variables

clc

str = fileread('22-Dec-2014_Sam-3.txt');

xpr = '(.+?)=(.+?),';

tkn = regexp(str,['\s*\S+:\s+(\S+)\s+',xpr,xpr,'([^\n]+)'],'tokens');

tkn = vertcat(tkn{:});

names = tkn(:,1);

idx = ~cellfun('isempty',strfind(tkn(:,1),'C-16'));

Psi = cellfun(@(s)sscanf(s,'%f'),tkn(:,3));

Delta = cellfun(@(s)sscanf(s,'%f'),tkn(:,5));

Soumya

[Stephen23]

For the second time: Please use the comments for commenting on other answers or your own question. The Answers are supposed to be for actually answering the question. Note that the order of the answer can change, so knowing what this "comment" applies to could be difficult in future.

[Stephen23]

The most robust and likely fastest method to access this data is to read all of Psi and Delta once into some numeric arrays, and then access whichever elements you need later using some indexing. This will be much faster when you need to match and access multiple different groups, and also much more robust for any future code changes and new data operations.

>> Psi(idx)

[Soumya Bhattacharya]

Hi, Thank you Stephen and sorry for the "comment" thing. Now the code is working great. Thanks a lot. Have a nice day. Soumya

[Soumya Bhattacharya]

Hi Stephen, Can you please help me understanding the code ? I cannot be able to understand the expression of 'xpr'(and it's role in next line) and the scanning format in next line and the syntax for 'cellfun'. Thank you.

[Stephen23]

The string

xpr = '(.+?)=(.+?),'

is used twice in the regular expression that is used in the function regexp. It matches any characters separated by an equals sign.

This will help you to understand cellfun:

http://se.mathworks.com/help/matlab/ref/cellfun.html