reading a file in matlab

hello,

I have a file like below named file.txt

> head(myfile,1:4])

              AT1G01060  AT1G01170   AT1G01260   AT1G01380
  AT1G01060  1.00000000  0.3885284 -0.14720327 -0.01865947
  AT1G01170  0.38852841  1.0000000 -0.29069241  0.26992353
  AT1G01260 -0.14720327 -0.2906924  1.00000000  0.30973373
  AT1G01380 -0.01865947  0.2699235  0.30973373  1.00000000
  AT1G01490  0.24681279  0.3955740 -0.07497821  0.23271890
  AT1G01500  0.05720335 -0.1786700 -0.26813919 -0.60440141

> dim(myfile)

[1] 2885 2885

please someone help me to read this file in matlab I was really exhausted

thank you

[Ken Atwell]

Can you clean up the formatting? I can't tell where the file content begin, or where line breaks may be.

[fereshteh izadi]

thank you, please help me...

[Image Analyst]

You forgot to attach file.txt. Make it easy for us to help you, not hard.

How did that file get created? Can you put a column header over the first row? If so, you can easily use readtable().

[fereshteh izadi]

this file is a correlation matrix derived from R function cor

cor(file.txt) file is too big to be attached but i attached the file i produced this file by cor function I should read this file then I could run this matlab code but I never could read the file

> head(mycounts[,1:6])

            AT1G01060  AT1G01170   AT1G01260   AT1G01380   AT1G01490   AT1G01500
AT1G01060  1.00000000  0.3885284 -0.14720327 -0.01865947  0.24681279  0.05720335
AT1G01170  0.38852841  1.0000000 -0.29069241  0.26992353  0.39557397 -0.17866996
AT1G01260 -0.14720327 -0.2906924  1.00000000  0.30973373 -0.07497821 -0.26813919
AT1G01380 -0.01865947  0.2699235  0.30973373  1.00000000  0.23271890 -0.60440141
AT1G01490  0.24681279  0.3955740 -0.07497821  0.23271890  1.00000000 -0.11413300
AT1G01500  0.05720335 -0.1786700 -0.26813919 -0.60440141 -0.11413300  1.00000000

[Walter Roberson]

In MATLAB, please do

filecontent = fileread('file.txt');
first20 = filecontent(1:20);
first20
double(first20)

and show us the output. This will allow us to figure out details of how the file is stored.

[fereshteh izadi]

thank you, i did like so

>> first20 = filecontent(1:20);

>> first20

first20 =

AT1G01060 AT1G01170

>> double(first20)

ans =

Columns 1 through 13

    65    84    49    71    48    49    48    54    48     9    65    84    49

Columns 14 through 20

    71    48    49    49    55    48     9

>>

[per isakson]

In your question the column and row headers are identical. However, in the sample file, tRMA.txt, the column and row headers are NOT identical. In what kind of variable do you want the headers? (ND.m indicates that the numerical data shall be transferred to a double array.)

[fereshteh izadi]

yes tRMA.txt is a transposed matrix of my expression data sets that by cor function I derived a correlation matrix with identical column and row headers as an input for ND.m matlab code but I can;t read my file sorry by " ND.m indicates that the numerical data shall be transferred to a double array", you mean ND.m needs a two dimential matrix without column or row name??? if so then after running the code on a numerical double array how i can assign column and row name to the output of ND.m????

[per isakson]

Yes, ND.m takes a square double array and no strings. Try the code in my answer. It should read files like tRMA.txt regardless of the number of columns and rows.

Answer 1

[per isakson]

• I failed to read tRMA.txt with Import Data. It choked Matlab (R2013a)
• The code below reads the file, tRMA.txt.
• I was a little surprised to see that num isn't square.

fid = fopen('tRMA.txt');
str = fgetl( fid );
[~] = fclose( fid );
colhead = strsplit( str, '\t' ); 
ncolumn = length( colhead );
fid = fopen('tRMA.txt');
fmt = ['%s',repmat('%f',[1,ncolumn])];
cac = textscan( fid, fmt, 'Headerlines',1, 'Delimiter','\t', 'CollectOutput',true );
[~] = fclose( fid );
rowhead = cac{1,1};
num     = cac{1,2};
whos colhead rowhead num

outputs

    Name           Size                Bytes  Class     Attributes
    colhead        1x2885             375050  cell                
    num          164x2885            3785120  double              
    rowhead      164x1                 22632  cell

&nbsp

In responce to comments:

I have converted the script to a function, preND. Functions are easier to use than scripts. Run

>> [ M, colhead, rowhead ] = preND( 'correlation.txt' );
>> mat_nd = ND( M );
>> imagesc( mat_nd )

where

function    [ M, colhead, rowhead ] = preND( filespec )
    fid = fopen( filespec );
    str = fgetl( fid );
    [~] = fclose( fid );
    colhead = strsplit( str, '\t' ); 
    ncolumn = length( colhead );
    fid = fopen( filespec );
    fmt = ['%s',repmat('%f',[1,ncolumn])];
    cac = textscan( fid, fmt, 'Headerlines',1               ... 
                ,   'Delimiter','\t', 'CollectOutput',true  );
    [~] = fclose( fid );
    rowhead = cac{1,1};
    M = cac{1,2};
end

the result is

&nbsp

The names of the rows and the columns are in the cell arrays, rowhead and colhead.

[fereshteh izadi]

thank you so much you both for your kindly attention. actually I should admit that I am too weak to adapt your code on my data and I got error. then I attached my file which ND.m code needs as input. as you mentioned already this code needs a double numerical array. my question is, how I can retrieve back my row names and colnames after running ND.m code????? you suppose I used the below syntax

M = dlmread('correlation.txt','\t')

mat_nd=ND(M,varargin)

it gives me an output but without rownames and colnames. then how i can have an output with the same header in my input correlation.txt??? please find my attached input file thanks again

[per isakson]

• "adapt your code on my data and I got error" &nbsp You make it difficult to help! You don't tell what you do or what error you encounter. I cannot guess! And why did you upload tRMA.txt?
• I cannot believe that &nbsp M = dlmread('correlation.txt','\t') &nbsp reads the file, which you uploaded together with your comment. Not with R2013a anyhow.

&nbsp

• Here &nbsp mat_nd=ND(M,varargin) returns

&nbsp

• "how I can retrieve back my row names and colnames after running ND.m code?????" &nbsp ND is totally ignorant of the row and column names; the names are neither in the input nor output of ND. Furthermore, dlmread cannot read the names - afaik.

[fereshteh izadi]

Dear per isakson, many thanks I performed as you commented but I got error by reading correlation.txt

let me ask an obvious question: how I can read correlation.txt in matlab??? I never could read this file in matlab and when i removed the colnames and rownames, by

M = dlmread('correlation.txt','\t')

i could read that...how i can read that when it is as i attached as correlation.zip???

actually I used a file without header as input for ND, I mean in R using rownames and colnames = FALSE, I removed the header of correlation.txt then i used as input. my question is; when ND gives me an output, is there anyway to back the header from the original correlation.txt(before removing the header) to the output ????

 *[ M, colhead, rowhead ] = preND( 'correlation.txt' );
Undefined function 'preND' for input arguments of type 'char'.*

[per isakson]

"how I can read correlation.txt in matlab???" &nbsp By the following steps

1. copy&paste the definition of the function, preND, from my answer to the Matlab editor (or any editor)
2. save the function to a file named, preND.m, in the folder where you have ND.m
3. put a copy of correlation.txt in the same folder
4. run [M,colhead,rowhead] = preND('correlation.txt');
5. if that fails, make sure that the folder is in the Matlab path.

"is there anyway to back the header from the original correlation.txt" &nbsp preND already did that (see my answer). However, how do you want to use all those names?

[fereshteh izadi]

thank you for your patience

>> pwd

ans =

C:\Users\Lenovo\Desktop\RMA

>> ls

.                        GENEI3.R                 RMA.txt                  completecorrelation.txt  evaluate.r               silencing.txt            
..                       Genes_in_GoldST.txt      SILENCING.m              correlation.txt          ggm.r                    silencing1.txt           
Ara_GoldST.txt           ND.m                     big.txt                  dream.txt                preND.m                  test.r                   

>> [M,colhead,rowhead] = preND('correlation.txt');

Undefined function 'strsplit' for input arguments of type 'char'.

Error in preND (line 6)
    colhead = strsplit( str, '\t' );

>>

[Walter Roberson]

strsplit was introduced in R2013a. Which MATLAB version are you using?

[fereshteh izadi]

 R2012a

[per isakson]

Replace

colhead = strsplit( str, '\t' );

by

colhead = regexp( str, '\t', 'split' );

[fereshteh izadi]

thank you very much, your code read correlation.txt well.

Answer 2

[Walter Roberson]

 fid = fopen('tRMA.txt, 'rt');

 %All columns are tab separated, but there is no initial tab before the first gene row
 %header which corresponds to the second column of input for the rest of the file,
 %with the first column of input being a row name string

 header = fgetl(fid);
 col_names = regexp(header, '\t\, 'split');
 num_cols = length(col_names);
 fmt = ['%s', repmat('\t%f', 1, num_cols)];

 datacell = textscan(fid, fmt, 'CollectOutput', 1, 'Delimiter', '\t');
 fclose(fid);

 row_names = datacell{1};
 cor = datacell{2};

Now there is row_names (a cell array of strings), col_names (a cell array of strings), and cor (a rectangular numeric matrix)

Answer 3

[A Mugesh]

Hi,

I am a beginer in mat lab learning, i want to know the code how to read the different format of files in matlab....