How can I add certain commands to .bash_history permanently? - bash

Question

I use bash and I sometimes have to type some very long commands for certain tasks. These long commands are not regularly used and generally gets overwritten in .bash_history file by the time I use them again. How can I add certain commands to .bash_history permanently?
Thanks.

Answer 1

As others have mentioned, the usual way to do that would be to store your long commands as either aliases or bash functions. A nice way to organise them would the to put them all in a file (say $HOME/.custom_funcs) then source it from .bashrc.
If you really really want to be able to load commands into your bash history, you can use the -r option of the history command.
From the man page:
-r Read the current history file and append its contents to the history list.
Just store all your entries in a file, and whenever you need your custom history loaded simply run history -r <your_file>.
Here's a demo:
[me#home]$ history | tail # see current history
1006 history | tail
1007 rm x
1008 vi .custom_history
1009 ls
1010 history | tail
1011 cd /var/log
1012 tail -f messages
1013 cd
1014 ls -al
1015 history | tail # see current history
[me#home]$ cat $HOME/.custom_history # content of custom history file
echo "hello world"
ls -al /home/stack/overflow
(cd /var/log/messages; wc -l *; cd -)
[me#home]$ history -r $HOME/.custom_history # load custom history
[me#home]$ history | tail # see updated history
1012 tail -f messages
1013 cd
1014 ls -al
1015 history | tail # see current history
1016 cat .custom_history
1017 history -r $HOME/.custom_history
1018 echo "hello world"
1019 ls -al /home/stack/overflow
1020 (cd /var/log/messages; wc -l *; cd -)
1021 history | tail # see updated history
Note how entries 1018-1020 weren't actually run but instead were loaded from the file.
At this point you can access them as you would normally using the history or ! commands, or the Ctrl+r shortcuts and the likes.

Answer 2

How about just extending the size of your bash history file with the shell variable
HISTFILESIZE
Instead of the default 500 make it something like 2000

Answer 3

The canonical answer is to create scripts containing these commands.
Edit Supposing you have the following history entries;
find /var/www -name '*.html' -exec fgrep '<title>' {} /dev/null \;
find ~/public_html -name '*.php' -exec fgrep include {} /dev/null \;
... you can try to isolate the parameters into a function something like this;
r () {
find "$1" -name "*.$2" -exec fgrep "$3" {} /dev/null \;
}
... which you could use like this, to repeat the history entries from above:
r /var/www html '<title>'
r ~/public_html php include
Obviously, the step is then not very long to create a proper script with defaults, parameter validation, etc. (Hint: you could usefully default to the current directory for the path, and no extension for the file name, and add options like --path and --ext to override the defaults when you want to; then there will be only one mandatory argument to the script.)
Typically, you would store the script in $HOME/bin and make sure this directory is added to your PATH from your .profile or similar. For functions, these are usually defined in .profile, or a separate file which is sourced from this file.
Having it in a central place also helps develop it further; for example, for precision and perhaps some minor added efficiency, you might want to add -type f to the find command; now there is only one place to remember to edit, and you will have it fixed for good.

Answer 4

You can create an alias for your command.
alias myalas='...a very long command here...'