[SOLVED] Using an ERE Pattern File for a grep Inversion Search

Issue

I need to use regex patterns in a grep pattern file (ERE) as BRE is insufficient. If data file contains:

alpha 1
bravo 1
gamma 1
delta 1
omicron 1
sigma 1
alpha 2
bravo 2
gamma 2
delta 2

and the grep inversion pattern file contains:

alpha
bravo
gamma
delta

then the following grep inversion: $ grep -v -f pattern_file main_file produces

omicron 1
sigma 1

This is the desired BRE output. But I need to tune the pattern file ERE so that the keywords stipulate beginning of line+keyword+space with the pattern file entries defined as:

"^alpha "
"^bravo "
"^gamma "
"^delta "

What is the correct grep inversion to produce the same two record output? The following does not work: $ grep -E -v -f pattern_file main_file

Solution

With the pattern file set as originally posted, a solution has been found that uses only a single grep inversion on an array loaded with the keywords from the main_file. No additional *nix tools used. Organizing ".bash_history" into ".bash_history.cmd" category files

declare -a keywords
keywords=(alpha bravo gamma delta)
IFS=\|; grep -E -v "^(${keywords[*]}) " main_file

which produces:

omicron 1
sigma 1

This satisfies the beginning of line anchor requirement as well as the space after the keyword, using an alternative token | to account for {keywords[*]} expansion, separated by the IFS character.

Answered By - Paul_Capo

Answer Checked By - Clifford M. (WPSolving Volunteer)