Facing problem with awk use

By default, fields are separated by runs of whitespace in AWK. It doesn't care how much white space there is, so a file like this would give the same result:

PROD   192.168.100.25   Unix   Active
PROD   192.168.100.26   Unix   -        Server   Active

As far as it's concerned, the fourth fields of those two lines are obviously Active and -.

Now, it's possible that your file format is actually such that the fields are separated by tabs, and only contain spaces in the middle. I don't think tabs survive posting in SE, and the spacing you show doesn't exactly match what you'd get with 8 columns wide tabs either, but I'll note this anyway.

Then, the lines would be more like

PROD<tab>192.168.100.26<tab>Unix - Server<tab><tab>Active

and you could tell AWK to use runs of tabs as separator like so:

$ awk -F '\t+' '{print $2, $4}' file.txt
192.168.100.25 Active
192.168.100.26 Active

Though, it'd be more common to have a single tab between each field, which would mean they wouldn't line up so nicely with variable-length data. Then you'd just use awk -F '\t'.

If the columns are actually fixed-width, you could use e.g. cut to pick the parts you need. This may involve manually counting the characters, though.

         1         2         3         4         5         6         7
1234567890123456789012345678901234567890123456789012345678901234567890
PROD   192.168.100.29   windows mail gateway       Active down 

Taking the count from above, this might work:

$ cut -c8-23,52-70 file-fixed.txt
192.168.100.25  Active
192.168.100.26  Active
192.168.100.27  Active
192.168.100.28  Not Active
192.168.100.29  Active down 
192.168.100.30  Active down

At least GNU AWK also has support for fixed-width fields, but I haven't looked too much into it. See https://www.gnu.org/software/gawk/manual/html_node/Fixed-width-data.html

First off, to be clear, awk is doing exactly what you asked it to do. By default, it will separate fields on any whitespace, and since the columns of interest can contain space characters, every space denotes a new field to awk.

When you have two arbitrary string inputs, it gets pretty hard to separate fields after they've been parsed by the linux column command. In my experience it is better to call a more robust programming language such as Python. I believe the Python CSV library has tools to infer the format of fixed-width fields.

If you could change the last field to only use the options "Active" "Inactive" "Down" then you could just use the following:

awk '{ print $2 $NF }' Inputfile

$NF means "number of fields" and is equivalent to selecting the last field.

But, the last column of data may or may not contain a space, which breaks this syntax.

Alternatively, as others have pointed out, if the fields are tab delimited, you should be able to use:

awk -F $'\t' '{ print $2 $4 }' Inputfile

The -F flag specifies the field separator, and $'\t' is the tab character.

However, if the entries are not tab delimited, you probably need to do something specific to this data format. If there are other input files that are formatted differently, you might still see failures.

Here's an example that will use any instance of two or more space characters as a field separator:

awk -F ' {2,}' '{ print $2 $NF }' Inputfile

In this case our field separator is a regular expression. It amounts to "any instance of two or more space characters."

This should work for the example you've given, but if you needed a field other than the first or last field, you could still run into trouble when column 3 contains multiple spaces.

To use awk in this case first replace the space in "Not Active" and also "Active down" by some other character:

sed 's/Not Active/Not_Active/g ; s/Active down/Active_down/g' inputfile

Then use awk and extract the second and last fields:

awk '{print $2,$NF}'

and finally restore those spaces:

sed 's/_//g'

Putting it all together the final command is:

sed 's/Not Active/Not_Active/g ; s/inputfile down/Active_down/g' active | awk '{print $2,$NF}' | sed 's/_/ /g'