output of top on a linux box

[Jeff McNeil]

The code snippet prints "cpu usage: us,, <type 'str'>" on my
workstation, bumping the splt index to 3 prints the correct sys.
usage.  What does the header of your 'top' command look like?  There's
a very good chance it's different than what I'm testing with.

As for the slice, assuming 'splt[2]' is '0.5%sy', then you should be
able to remove '%sy' using an indexof -3:

Python 2.4.3 (#1, Oct 23 2006, 14:19:47)
[GCC 4.1.1 20060525 (Red Hat 4.1.1-1)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> a = "0.5%sy"
>>> a[:-3]
'0.5'
>>>

Also note that you may very well have some curses/terminal chars
tagged on to that '0.5%sy' string, which causes the slice to return
something that you're not expecting.  For example:

Python 2.4.3 (#1, Oct 23 2006, 14:19:47)
[GCC 4.1.1 20060525 (Red Hat 4.1.1-1)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import os
>>> top = os.popen("top -n 1")
>>> rd = top.read().split("\n")
>>> splt = rd[2].split()
>>> print splt
['Cpu(s):\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m', ' 1.2%',
'\x1b(B\x1b[m\x1b[39;49mus,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
'0.2%', '\x1b(B\x1b[m\x1b[39;49msy,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
'0.1%', '\x1b(B\x1b[m\x1b[39;49mni,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
' 98.4%', '\x1b(B\x1b[m\x1b[39;49mid,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
'0.1%', '\x1b(B\x1b[m\x1b[39;49mwa,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
'0.0%', '\x1b(B\x1b[m\x1b[39;49mhi,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
' 0.0%', '\x1b(B\x1b[m\x1b[39;49msi,\x1b(B\x1b[m\x1b[39;49m\x1b(B\x1b[m',
'0\x1b(B\x1b[m\x1b[39;49m\x1b[K']
>>>

Jeff


Are you sure that splt only contains text? Are there any terminal
control characters?
On 4/9/07, Pradnyesh Sawant <spradml at gmail.com> wrote:
> Thanks for the pointer to use /proc/stat
> However, that does not answer my other question about string slicing
> Any pointers in that direction would be helpful...
>
> Thanks a lot!
>
> On 4/9/07, Jeff McNeil <jeff at jmcneil.net> wrote:
> > Can you pull the same information from /proc/stat as opposed to using
> > a pipe to top? The first line(s) should contain (at least):
> >
> > cpu usermode, lowprio, system, idle, hz.
> >
> > The 2.6 kernel adds iowait, irq, and soft irq.  It seems that this
> > might be a better solution than executing that additional command.
> >
> > Take a look at proc(5).
> >
> > Thanks,
> >
> > Jeff
> >
> >
> >
> >
> >
> > On 4/9/07, Pradnyesh Sawant <spradml at gmail.com> wrote:
> > > Hello,
> > > I need the cpu usage of a linux box, for which i capture the output of
> > > "top" using "popen". However, i am facing problems during string
> > > handling. The code snippet is:-
> > >
> > > top = os.popen("top -n 1")
> > > rd = top.read().split("\n")
> > > splt = rd[2].split()
> > > # cpu = splt[2][:-4]  # why ain't this working ???
> > > cpu = splt[2]
> > > print "cpu usage: %s, %s" % (cpu, type(cpu))
> > >
> > > the "print" statement prints, "cpu usage: 0.5%sy, <type 'str'>". i
> > > tried my best to get rid of the "%sy" after the "0.5", but couldn't
> > > succeed (even though the type of "cpu" is "string").
> > >
> > > Also, one weird thing that was happening was that if i still "split"
> > > the cpu string (removed the last 4 characters), the output of the next
> > > print statement was wrong (specifically, the first 2 characters got
> > > chipped off). This output is totally unexpected, and wrong...
> > >
> > > Can anyone kindly help me in understanding what exactly is going on here.
> > >
> > > Thanks a lot!
> > > --
> > > http://mail.python.org/mailman/listinfo/python-list
> > >
> >
>