Clearing commands list, and other tales of bash history

Bash command history file

Bash saves a list of the commands run in a session into the .bash_history file in the home directory. You can view its contents by symply typing in the terminal

❯ history

Btw, another way to clear the history is simply to delete this file.

Delete just a command, or a range of commands

Besides the -c option to delete the complete history, there is also a the -s to delete specific entries in the list

# whole history
❯ history -c

# command 123 in the history
❯ history -s 123

# commands 123, 124. , 125
❯ history -s 123:125

N and M can be absolute numbers or relative. The absolute numbers you can get by running history, which prints them out next to the command. As for the relative ones I have never needed the extra complexity so I avoid them. Talking of complex edge case I’ve never needed, you can also write to the history file with a few different variations. Here’s a good reference from the GNU manual

Cycling through only the commands starting with a certain string

If you press the up arrow key at the start of a command line prompt, bash will cycle backwards through all the commands you typed previously. Pressing the up arrow cycles through all the commands, which can be tedious if the one you are after is far away. But if you know how the command started, you can type the start and then the up / down arrow to limit your search only to the commands starting with that string. To enable that you need to add these two lines to your .inputrc file (if the file doesn’t exist, just create it):

"\e[B": history-search-forward
"\e[A": history-search-backward

Searching through the command history

A more powerful way is to use reverse-i-search. The i stands for “incremental”, because bash shows you a result as soon as you type a character, and then it refines the search as you type the next character, and so on. A typical incremental search goes like this

1. Enter <CTRL>-R on the terminal. That will bring up the prompt (reverse-i-search)`':
2. Type a few characters from the commands you are after
3. As you type, bash will bring up the most recent command that matches it
4. You can cycle backwards through the list of matching commands by entering more <CTRL-R>. In theory you could use <CTRL-S> to cycle forward, but that’s not enabled on many terminals. I jave never needed it enough to make me want to research how to enable it, but here’s a starting point on Stack Overflow if you are interested.
5. Once you find the command you can
   1. Press <ENTER> to run the command again
   2. Press right arrow ➡️ to enter in the command and edit
   3. Press <CTRL>-G to exit the search without running the command

Prevent certain commands from getting into the history in the first place

It can also be useful to prevent certain commands from getting into the history in the first place. There are two different approaches.

To keep your history file size down, you can ignore sequential duplicates with the export HISTCONTROL=ignoredups line in your .bash_profile. Another trick, to allow you to bail out of history on a command by command basis, is to ignore commands that start with a blank character. That is achieved with export HISTCONTROL=ignorespace. Then if you want to run a command without it being recorded in the history, just type a blank space and then the command. If you want both options, combine them with a colon - mine, for example is export HISTCONTROL=ignoredups:ignorespace

Another useful setting is a list of commands (wildcards work) you don’t want to go in the history. These are typically generic commands that don’t add any value to the history, like looking up a man page, or, indeed, history. The variable to set is HISTIGNORE, and again, the separator is the colon. For example mine is export HISTIGNORE="ls:cd:cd -:pwd:exit:date:* --help:man *:la:history"

Change how many commands bash remembers in its history

The number of commands saved by bash is controlled by the environmental variable HISTSIZE, which is by default 500. You can add export HISTSIZE=XXX, where XXX is another number, as a new line to .bash_profile if you want to change that. I have a large one, 32768.

Don’t use a history file at all

Or you can avoid having a history file at all by deleting HISTSIZE altogether. That can be achieved by adding the line unset HISTSIZE to your .bash_profile. Or alternatively, setting it to 0, it does the same thing. This may be useful on shared server environments.

Summary: all the options mentioned in the post

Here are the options mentioned in the post. This would be added to your .bash_profile or .bashrc in your home directory.

# no history
export HISTSIZE=0
# or
unset HISTSIZE

# larger history than default
export HISTSIZE=32768

# ignore sequential duplicates
export HISTCONTROL=ignoredups

# ignore lines that start with a space
export HISTCONTROL=ignoredspace

# both of the above
export HISTCONTROL=ignoredups:ignorespace

# don't put the commands between the colons in the history
export HISTIGNORE="ls:cd:cd -:pwd:exit:date:* --help:man *:la:history"