How to add a custom string to a Matlab script's output?

Dear community,

I have the following issue to solve. I run some data processing in a cluster environment, parallelized both across nodes and within a node with the use of parfor. The output is dumped into a single text file and there is no way to distinguish which line belongs to the processing of which data set. I can separate output from different nodes into individual log files, but not from different parallel jobs on a node, since it is treated by the cluster as a single job. My Matlab code is basically a wrapper around the data processing Matlab script. Is it possible to somehow change each output line of a Matlab script and add a specific string to the beginning of it, e.g. a dataset ID? The only solution I see is to go through the script, hunt down lines that output something and append something to it. Basically, I am looking for something like this in a bash shell:

script_name | sed 's/^/$DATA_ID/'

Best regards,

Renat.

[Rik]

You could replace all occurrences of fprintf with your own function that appends your string and then calls fprintf again to write it. I would suggest you prepend the ID instead of append, as the fprintf call is likely to end with some form of a newline.

[Renat]

Well, I was hoping there is a way to avoid doing this.

Answer 1

[Swastik Sarkar]

Hi @Renat,

It is possible to override the built-in function responsible for displaying data to achieve the desired prefix on each output on each worker. For example, if the disp built-in function is used to display data, the following function can be defined to override it:

disp('Output')

prefix: Output

function disp(arg)
    builtin('disp', ['prefix: ' arg])
end

For more information on the builtin function, refer to the following documentation:

https://www.mathworks.com/help/matlab/ref/builtin.html

Hope this helps printing desired info without requiring extensive modifications to the entire code.