Re: use of parallel in bash script

[Dietmar Pils]

dear ole,

if you have read (http://en.wikipedia.org/wiki/Reading_%28process%29)
the problem accurately you would have recognised, that it is a simple
problem, not the command per se doesn't work, but it doesn't work from
inside a script: it seems that you have to export the command (this is
not very well documented). this you could have advised without
instructing users how to write error reports...

dietmar



On 12/17/2012 01:23 AM, Ole Tange wrote:
> On Thu, Dec 6, 2012 at 9:04 PM, Dietmar Pils <dietmar.pils@univie.ac.at> 
> wrote:
>
>> how can I use parallel in a bash script? what do I miss, but the snippet
>>
>> time cat RUM.sam | parallel --pipe --files -N 5184953 sort -k 3,3 -k 4,4n |
>> parallel -mj1 sort -k 3,3 -k 4,4n -m {} ";" rm {} > sort.sam
>>
>> works from the command lane but not in a bash script.
>>
>> what have I to do, that I can run the same from inside a bash script...
> You need to learn to write better error reports. See for instance:
>
> http://www.ehow.com/how_6946968_write-error-report.html
>
> One thing that article leaves out is: Try to minimize/simplify your
> problem. If steps A B C D do not work, you need to let the programmer
> know if A, A B, and A B C do work.
>
>
> /Ole
>
> PS: sort from GNU coreutils 8.6 uses multiple CPUs for sorting, so
> using GNU Parallel for sorting is less efficient.