SYNOPSIS From script: use Bash::History::Read qw(parse_bash_history_file); my $res = parse_bash_history_file("$ENV{HOME}/.bash_history"); Sample result: [ [undef, "some-command\n"], [1446715184, "du -sm\n"], [1446715190, "ls -l\n"], ] From the command-line: % perl -MBash::History::Read -i.bak -e'each_hist { $PRINT = 0 if $TS < time()-2*30*86400; # delete old entries $PRINT = 0 if /foo/; # delete unwanted lines (e.g. matching some regex) s/(mysql\s+-p)(\S+)/$1******/; # redact sensitive information }' ~/.bash_history DESCRIPTION This module provides utility routines to read entries from bash history file (by default ~/.bash_history). The format of the history file is dead simple: one line per entry, but when HISTTIMEFORMAT environment is set, bash will print a timestamp line before each entry, e.g.: #1374290613 ls -al #1374290618 less myfile #1374290635 ... See each_hist for one routine to let you handle this format conveniently. FUNCTIONS parse_bash_history_file([ $path ]) => array Parse entries from bash history file. If unspecified, $path will default to HISTFILE environment variable or $HOME/.bash_history. Return an array of entries, where each entry is <[$timestamp, $line]> and $timestamp can be undef if entry does not have a timestamp. each_hist { PERL_CODE } Will read lines from the diamond operator (<>) and call Perl code for each history entry. Can handle timestamp lines. This routine is exported by default and is meant to be used from one-liners. Inside the Perl code, $_ is locally set to the entry content, $TS is locally set to the timestamp (and changes to this variable is ignored, except when you undefine the variable, which will remove the timestamp from output), $PRINT is locally set to 1. If $PRINT is still true by the time the Perl code ends, the entry (along with its timestamp) will be printed. So to remove a line, you can set $PRINT to 0 in your code. To modify content, modify the $_ variable.