I have an array which looks like this:
cloud =
7.5059 51.4406
7.5057 51.4445
7.5048 51.4484
7.5034 51.4522
7.5014 51.4558
7.4989 51.4593
7.4958 51.4627
7.4923 51.4658
7.4884 51.4686
.
.
all i want is to write this array to a text file as it is, in the same format. I tried both fprintf and dlmwritebut i'm able to produce the exact same format. I know its an easy one, but I'm only asking after trying a lot.
Have you looked into string formatting?
fid = fopen( 'myFile.txt', 'w' );
for ii=1:size(cloud,1)
fprintf( fid, '%.5g\t%.5g\r\n', cloud(ii,1), cloud(ii,2) );
end
fclose( fid ); % do not forget to close the file :-)
Have you considered saveing into ascii file?
save( 'myFile.txt', 'cloud', '-ascii', '-tabs' );
EDIT:
End-of-line issue: for text file there are several way of marking the end of line: On windows it is usually required to print \r\n, for Mac and Linux sometimes it is enough to use \r and sometimes \n (I'm not 100% sure). So, you might need to experiment with it a bit to find out what works best for your machine. (Thanks #Rody for correcting me here)
Accuracy: the number in the formatting string %.5g determines the accuracy of the printed number. Again, you can play with it till you are satisfied with the results.
WINDOWS
Here's one way:
fid = fopen('cloud.txt', 'w');
fprintf(fid, '%.4f\t%.4f\r\n', cloud.');
fclose(fid)
Here's the more readable way:
dlmwrite('cloud.txt', cloud, 'delimiter', '\t', 'precision', 4, 'newline', 'pc')
LINUX
Here's one way:
fid = fopen('cloud.txt', 'w');
fprintf(fid, '%.4f\t%.4f\n', cloud.');
fclose(fid)
Here's the more readable way:
dlmwrite('cloud.txt', cloud, 'delimiter', '\t', 'precision', 4);
Related
Neither of the following line write "a\b" into file
fid = fopen('myfile.txt','w'); fprintf(fid, 'a\b'); fclose(fid);
fid = fopen('myfile.txt','wb'); fprintf(fid, 'a\b'); fclose(fid);
Probably, Matlab does backslash escaping during save to file.
How to disable this "feature"?
The string should remain intact, i.e. fprintf(fid, 'a\\b') is not a solution, length('a\b')==3 should be true.
You could use conversion characters in fprintf, i.e. %s in this case
fid = fopen('myfile.txt','w'); fprintf(fid, '%s', 'a\b'); fclose(fid);
Your condition length('a\b')==3 isn't violated like this, if I am not mistaken
Another alternative is to use fwrite:
fwrite(fid,'a\b','uchar');
If you use 'char' mode, the string will be encoded differently depending on how the file was opened (e.g. UTF encoding).
I have a huge .txt file and parts of which I want to parse (using text scan), say I have 10000 line data and a part which starts at line 300, the part also has a header of 10 lines say,how can I skip the first 300 lines(not using header function of text scan of course as I then wont be able to get my actual 10 line header) or is there a way in which I can jump to line 300 and start text scan from there as if 301 line was the first line.
So, assuming your data is generated by the following (since you have not mentioned how it's formatted in your question):
fid = fopen('datafile.txt', 'w');
for i=1:300
fprintf(fid, 'ignore%d\n', i);
end
for i=301:310
fprintf(fid, 'header%d\n', i);
end
for i=311:10000
fprintf(fid, '%d\n', i);
end
fclose(fid);
You can read the data by first using fgetl to advance to the 300th line, then using textscan twice to get the header info and the data. Basically, the thing to remember is that textscan works starting from the place where fid is pointing. So, if it's pointing to the 301st line, it'll start scanning from there. So, here's the code to read the file above, starting from line 301:
fid = fopen('datafile.txt', 'r');
for i=1:300
fgetl(fid);
end
scannedHeader = textscan(fid, '%s', 10);
scannedData = textscan(fid, '%d');
fclose(fid);
NB: if the data is always the same format, you can use ftell to know where to skip to exactly then use fseek to go to that offset.
I am trying to read a csv file use textscan. The fields are seperated with ',' . I used the following code, but it only read in one line of data into the matrix W.
I also tried dlmread(), it got the number of fields wrong.
The file is contructed under linux, matlab is under linux.
file_id = fopen('H:\data\overlapmatrices\cos.mat.10');
W = textscan(file_id, '%f', 'delimiter', ',' , 'EndOfLine', '\r\n');
fclose(file_id);
clear file_id;
you might wanna try csvread, it should do the trick.
or you could alway do something dirty like
fid = fopen( filename );
tline = fgetl(fid);
while ischar(tline) %or some other check
%sscanf(tline...
tline = fgetl(fid);
end
The problem could be in how the end of line is represented in the file (see also this article on Wikipedia). While \r\n (the combination of a carriage return and a newline character) is common on Windows, \n (just the newline character) is the standard on Linux and other Unix systems.
But as ben is saying, csvread might be an easier way how to read the file.
I want to load a csv file in a matrix using matlab.
I used the following code:
formatSpec = ['%*f', repmat('%f',1,20)];
fid = fopen(filename);
X = textscan(fid, formatSpec, 'Delimiter', ',', 'CollectOutput', 1);
fclose(fid);
X = X{1};
The csv file has 1000 rows and 21 columns.
However, the matrix X generated has 2000 columns and 20 columns.
I tried using different delimiters like '\t' or '\n', but it doesn't change.
When I displayed X, I noticed that it displayed the correct csv file but with extra rows of zeros every 2 rows.
I also tried adding the 'HeaderLines' parameters:
`X = textscan(fid, formatSpec1, 'Delimiter', '\n', 'CollectOutput', 1, 'HeaderLines', 1);`
but this time, the result is an empty matrix.
Am I missing something?
EDIT: #horchler
I could read with no problem the 'test.csv' file.
There is no extra comma at the end of each row. I generated my csv file with a python script: I read the rows of another csv file, modified these (selecting some of them and doing arithmetic operations on them) and wrote the new rows on another csv file. In order to do this, I converted each element of the first csv file into floats...
New Edit:
Reading the textscan documentation more carefully, I think the problem is that my input file is neither a textfile nor a str, but a file containing floats
EDIT: three lines from the file
0,1,0,0,0,1,0,0,0,1,0,0,0,1,0,0,1,0,0,0,2
1,-0.3834323,-1.92452324171,-1.2453254094,0.43455627857,-0.24571121,0.4340657,1,1,0,0,0,0.3517396202,1,0,0,0.3558122164,0.2936975319,0.4105696144,0,1,0
-0.78676,-1.09767,0.765554578,0.76579043,0.76,1,0,0,323124.235998,1,0,0,0,1,0,0,1,0,0,0,2
How about using regex ?
X=[];
fid = fopen(filename);
while 1
fl = fgetl(fid);
if ~ischar(fl), break, end
r =regexp(fl,'([-]*\d+[.]*\d*)','match');
r=r(1:21); % because your line 2nd is somehow having 22 elements,
% all lines must have same # elements or an error will be thrown
% Error: CAT arguments dimensions are not consistent.
X=[X;r];
end
fclose(fid);
Using csvread to read a csv file seems a good option. However, I also tend to read csv files with textscan as files are sometimes badly written. Having more options to read them is therefore necessary.
I face a reading problem like yours when I think the file is written a certain way but it is actually written another way. To debug it I use fgetl and print, for each line read, both the output of fgetl and its double version (see the example below). Examining the double version, you may find which character causes a problem.
In your case, I would first look at multiple occurrences of delimiters (',' and '\t') and , in 'textscan', I would activate the option 'MultipleDelimsAsOne' (while turning off 'CollectOutput').
fid = fopen(filename);
tline = fgetl(fid);
while ischar(tline)
disp(tline);
double(tline)
pause;
tline = fgetl(fid);
end
fclose(fid);
I am using the following functions for writing and reading 4098 floating point numbers in MATLAB:
Writing:
fid = fopen(completepath, 'w');
fprintf(fid, '%1.30f\r\n', y)
Reading:
data = textread(completepath, '%f', 4098);
where y contains 4098 numbers. I now want to write and read 3 strings at the end of this data. How do I read two different datatypes? Please help me. Thanks in advance.
Here's an example of what I think you want to do, using TEXTSCAN for reading the file instead of TEXTREAD (which will be removed in a future version of MATLAB):
%# Writing to the file:
fid = fopen(completepath,'w'); %# Open the file
fprintf(fid,'%1.30f\r\n',y); %# Write the data
fprintf(fid,'Hello\r\n'); %# Write string 1
fprintf(fid,'there\r\n'); %# Write string 2
fprintf(fid,'world!\r\n'); %# Write string 3
fclose(fid); %# Close the file
%# Reading from the file:
fid = fopen(completepath,'r'); %# Open the file
data = textscan(fid,'%f',4098); %# Read the data
stringData = textscan(fid,'%s',3); %# Read the strings
fclose(fid); %# Close the file
Well, you can write out a string at any point when you are writing to the file with the following:
fprintf(fid, '%s', mystring);
Of course, you might want something more like the form you gave:
fprintf(fid,'%s\r\n', mystring);
And you could mix the floating point with the string like so:
fprintf(fid, '%1.30f %s\r\n', y, mystring);
If you are dealing with mixed data types, you might want to use fscanf instead of textread if the formatting isn't very regular. For instance,
data = fscanf(fid, '%s', 1);
reads one character string from the file.
Take a look at the help files for fscanf for more information on how to use it. These functions are pretty much ANSI C functions (fprintf and fscanf I mean) so you find more info on the web about them quite easily.