Regex pattern matching

[termvrl]

Hi All,

I tried to use regex to do a pattern matching for CPU % usage. i only need the value, 0.0.

The command i use:

top -bn1 | grep "Cpu(s)" | grep ".\.." | grep ".\..%us"

The result:

Cpu(s):  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st

After some time, i managed to get the right regex for the pattern matching but its not work in my server.
The error prompt out:

top -bn1 | grep \d.\d(?=%us)
-bash: syntax error near unexpected token `('

Any way i can solve this?
Thanks

[MG&TL]

For your error, that's because your shell (bash) attempts to parse the expression you give to grep, and fails. To avoid that, use quotes:

top -bn1 | grep '\d.\d(?=%us)'

However, that doesn't seem to work, at least not on my system, and I don't know enough about grep's pattern matching to fix it. Instead, I used a combination of grep (to get the right line), awk (to get the right column), and sed (trim excess rubbish):

$ top -bn1 | grep 'Cpu(s)' | awk '{print $2}' | sed 's/%us,//g'
7.4

[vasa1]

In this specific instance

top -bn1 | awk 'NR==3 { print $2}'

seems to do as well.

[steeldriver]

You are trying to use a perl-style regular expression - grep supports that to some extent, but you need to add the -P switch

$ top -bn1 | grep -P '\d.\d(?=%us)'
Cpu(s):  7.5%us,  5.0%sy,  0.0%ni, 87.2%id,  0.2%wa,  0.0%hi,  0.0%si,  0.0%st

If you want just the matching number, add -o

$ top -bn1 | grep -oP '\d.\d(?=%us)'
7.5

Note that '\d.\d' will match 2 digits separated by ANY single character - if you want to match just the decimal point, you should escape the period, either '\d\.\d' or '\d[.]\d'

[termvrl]

Hi,

Thanks for the response.
It works.

I will find more into 'awk' , 'sed' . its usefull.

[termvrl]

Hi steeldriver,

Thanks for your explanation.
I dont know that regex has different style. Thanks again.

[vasa1]

...
If you want just the matching number, add -o

top -bn1 | grep -oP  '\d.\d(?=%us)'

...

That code doesn't work for me but this does:

top -bn1 | grep -oP '\d\.\d +(?=us)'

I tried both on a regular desktop Xubuntu.

When I checked man pcre, I found this:

Lookahead assertions

Lookahead assertions start with `(?=` for positive assertions and `(?!` for
negative assertions. For example,

`\w+(?)`

matches a word followed by a semicolon, but does not include the semi‐
colon in the match, ...

This is my grep:

[12:56 PM] ~ $ grep --version
grep (GNU grep) 2.14
Copyright (C) 2012 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

Written by Mike Haertel and others, see <http://git.sv.gnu.org/cgit/grep.git/tree/AUTHORS>.

[steeldriver]

@vasa1 your expression matches a different pattern (digit-period-digit followed by one or more spaces) - is your 'top' output different i.e. '7.5 %us' instead of '7.5%us'?

FWIW for matching a floating point number in general I'd suggest something like

grep -oP '[\d.]+(?=%us)'

which will match a wider range of possible field widths and precisions i.e. 7%us, 7.5%us, 17%us, 27.5%us etc. (note that the period is treated as literal when inside a [ ] character set, so no need to escape it). If you need to allow for optional trailing whitespace beween the number and the % sign, you could add \s* i.e.

grep -oP '[\d.]+\s*(?=%us)'

or maybe more perlish

grep -oP '[\d.]+\s+?(?=%us)'

[vasa1]

This is what I get when I run top -bn1

[06:06 PM] ~ $ top -bn1
top - 18:06:09 up  2:27,  2 users,  load average: 0.04, 0.04, 0.05
Tasks: 133 total,   1 running, 132 sleeping,   0 stopped,   0 zombie
%Cpu(s):  2.7 us,  0.6 sy,  0.0 ni, 95.5 id,  1.2 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:   4010624 total,  1106888 used,  2903736 free,    46644 buffers
KiB Swap:  6159668 total,        0 used,  6159668 free,   660436 cached

So our top outputs do differ. Any idea why? My top is whatever came with the system.

[steeldriver]

I don't know - top has *lots* of configuration options (plus optional global / personal rc files)... or it could just be the top version? mine is

$ top --version
    top: procps version 3.2.8