I sometimes process files where I would like to be able to allow block 
boundaries to only occur at rows where "-block-breaks" occur.

I would like to be able to define such breaks as a perl expression, evaluated 
for each line, whose value must be different from the prior line for that line 
to be the beginning of a new block.

The expression should be able to refer to columns either by number or by 
-header name.

For example, I have a program to emits a graph for every protein, where every 
line is residue of the protein, and there is a column, proteinID, whose value 
changes when the protein changes which I would like to call as follows:

parallel -j 40 -cat -block 10K --block-breaks proteinID

In the meantime I suppose a workaround is to preprocess the input and insert 
fake -recstart wherever the column changes value.

Other workaround welcome....

Cheers,

Malcolm Cook

Reply via email to