Simple Python equivalent for the shell command

[Michael Torrie]

On 11/28/2016 08:08 AM, Ganesh Pal wrote:
> I  was trying  to write a function that will return me the unique number
> associated with each employee id.The command has the output in the below
> pattern
> 
> Linux-Box-1# employee_details ls
> List of names:
> 100910bd9 s7018
> 100d60003 s7019
> 110610bd3 s7020
> 100d60002 s7021
> 
> 
> Linux-Box-1# employee_details ls | grep "s7020" | awk '{print $1}'
> 100d60003
> 
> It's a one liner in Shell  :)
> 
> 
> I tried converting  the same  in the python style , Any better suggestion
> and loop holes in the below program

Well Bash is really good at some things.  Piping commands together is
one of those things.  Python can do such things but not in as compact a
way.  For one Python has no quick way of interfacing with subprograms as
if they were language-level constructs like Bash does.  But it still can
do a lot of the same tasks that shell scripting does.  Read on.

In many respects, generators are the Python equivalent of pipes. See
this: http://www.dabeaz.com/generators/, specifically
http://www.dabeaz.com/generators/Generators.pdf

Here's a simple grep-like filter that searches lines:

def simple_grep(search_in, look_for):
    for line in search_in:
        if look_for in line:
            yield line

import sys
for result in simple_grep(sys.stdin, "s2070"):
    print result.split()[0]

This obviously does not run a subprogram, but could feed the filter
input from any file-like object that returns lines.  The beauty of
generators is that you can string them together just like you would do
in Bash with pipes.  And they can potentially be fairly fast, especially
if you can reduce the number of lines the generators have to process by
putting the faster filters earlier on in the chain (much like you would
optimize a database query by pushing selections ahead of joins).  Though
this is not shorter than your code, nor is it substantially different,
it does have the advantage that the simple_grep() generator can be used
in multiple places and could even be placed in utility library.