Bash script Issue parsing text in line with whitespace characters

Question

I have a list of commands to parse through for an audit, similar to this:

1. -a *policy name=PolicyName -a *policy workflow name=PolicyWorkflow -a *policy action name=PolicyAction -s Server -b Storage -J Node -y 1 Months -o -F -S
2. -a *policy name=PolicyName -a *policy workflow name=PolicyWorkflow -a *policy action name=PolicyAction -s Server -b Storage -J Node -y 1 Months -o -F -S
3. -a *policy name=Policy Name -a *policy workflow name=PolicyWorkflow -a *policy action name=PolicyAction -s Server -b Storage -J Node -y 1 Months -o -F -S

I'm trying to set each of the name=Value to variables. As this is a standard pattern, i used read with success, however when I get to a line with whitespace (see line 3), it offsets all my variables. I'm unsure how to tackle this without looping through each word in the line and matching specific patterns. Hoping someone would have a better solution out there.

You could just use sed to replace all instances of 'Policy Name' with PolicyName prior to using read, e.g. `sed 's/Policy Name/PolicyName/g' | while read ...` — Warwick, May 26 '18 at 05:30
@Warwick Only if the whitespace is unimportant and the policy name is actually "Policy Name" and not just a placeholder for an actual policy name that the poster inserted to show the general pattern. — Kusalananda, May 26 '18 at 06:40

αғsнιη · Accepted Answer · 2018-05-28T11:23:14.080

In bash and using an array variable instead, you can do something like:

{ IFS=$'\n'; array=($(grep -Po 'name=[^-]+(?=\s*-)' infile)); }

then print the elements of the array (array Index in bash starts from 0):

printf '%s\n' "${array[@]}"
name=PolicyName
name=PolicyWorkflow
name=PolicyAction
name=PolicyName
name=PolicyWorkflow
name=PolicyAction
name=Policy Name
name=PolicyWorkflow
name=PolicyAction

Or to print just a single element:

printf '%s\n' "${array[6]}"
name=Policy Name

We defined IFS to "\newline" character IFS=$'\n' so that the word splitting will happen on "\newline" character only on the result of unquoted–Command Substitution $(...).
The syntax { list; } is known Grouping Commands and I just used to grouping my commands.
The array=(...) syntax is used to creating an associated array variable named array.
With grep -Po 'name=[^-]+(?=\s*-)' infile, we print -only the matches that matches the pattern "name=" followed by any non-zero length of characters but a hyphen [^-]+ which follows by zero-or-more whitespaces until a hyphen (?=\s*-).
- The syntax (?=...) is positive-lookahead and with grep's -P option it's can be use to interpret the pattern as a Perl-Compatible Regular Expression (PCRE).

Future reading:

@JustinDavis I added some explanations if you were interested to know what is happening. — αғsнιη, May 28 '18 at 11:18

Bash script Issue parsing text in line with whitespace characters

1 Answers1