Text file data segregation

I want to segregate a text file which contains data. Segregation should be done on the basis of asterix (*) in the text file. So the data should be read into one variable and when there is asterix it should start reading the data in another variable.

Eg.

**DATA1

.........................................

.........................................

.........................................

.........................................

**DATA2

........................................

..........................................

.........................................

..........................................

**DATA3

.............................................

...........................................

............................................

...................................

......................................

Contents below **DATA1 should be stored in DATA1 and should be there in workspace to view similarly contents under DATA2 should be visible. The code should automatically read that there is double ** and start reading the next data into that variable.

[Rik]

You should not use numbered variables. Use a cell array instead.

If you read your file as text it is easy to loop through the lines. What have you tried so far?

[MANISH R]

I have tried to segregate it based on the number of lines using textscan command with header lines and it works but the number of lines in all the files are not same so i have to make a generic one.

Answer 1

[Chunru]

type test.txt

**DATA1
10 
20
30
.**DATA2
40
50
**DATA3
60

result = {};
fid = fopen("test.txt", "rt");
while ~feof(fid)
    i = fscanf(fid, "**DATA%d");
    result{i} = fscanf(fid, "%f")';
end
result

result = 1×3 cell array
    {[10 20 30]}    {[40 50]}    {[60]}

[MANISH R]

@Chunru Thank you but this is little away from what was required by me. Result Im looking for is, if i click DATA1 in workbench then i should get

10

20

30

if i click DATA2 then

40

50

then DATA3

60

all separately.

Anyways thanks for the reply

[Chunru]

You are strongly encouraged to use cell array instead of different variable names such as DATA1, DATA2, and so on.

Answer 2

[Walter Roberson]

Use fileread() to read the file as a character vector. Use

parts = regexp(TheVector, '\*\*', 'split');
parts(cellfun(@isempty, parts)) = [];

Now parts is a cell array of character vectors. The ** has been removed, so the sections would start immediately with the DATA characters. Process each section as is appropriate for your purposes.

For example it might make sense to use textscan() of the text with 'headerlines', 1, and with format '' (empty string). When you use textscan() with empty string as the format, and the data is numeric columns, then textscan will automatically figure out how many columns are present.

That said, if you fopen() the file and you ask to textscan() the fid with '' format, then it should stop reading at each point it encounters the ** since that is not something that can be interpreted as numeric. You would then fgetl() and save the variable name, and then you would run another textscan() which would pick up from after that line.

[MANISH R]

@Walter Roberson Thank you. Splitting the data based on ** works fine but Im not able to write and store them as required. With textscan , headerlines it becomes dependent on number of lines which is not ideal because each file has different number of lines. It would be helpful if you could ellaborate little more as Im new to matlab coding. Thanks