More of a curiosity of if and how it can be done. I sometimes find myself repeating commands with small parts of the same command modified.

Example:

grep aaa file.txt
grep bbb file.txt
grep ccc file.txt

I know I could set this up with a for loop but I don't always know all of the values that may need to be changed/tested at start. I can also use keyboard movement to get back to the aaa quicker, but it's still extra keys. Not a lot with the example, but some strung together commands can result in a lot of keyboard movement needed.

I know that variables and files can be passed to certain commands at the end, for example.

while read i; do echo "$i"; done < ./file.txt

x='./file.txt'
sed 's/find/replace/' <<<${x}

I'm curious if there is a way with bash substitution to do similar declaration at the end of the command for undeclared variables. I tried variations like grep ${} file.txt <<< 'aaa' but so far nothing I've toyed with has worked out. My goal is that the passed value is not declared as a variable so I can just hit up arrow on keyboard and edit the string at the end of the command statement. Appreciate any help or insights that can be given!

CodePudding user response：

You can define a quick function:

f () { grep "$1" file.txt; }

Now you can run

f aaa
f bbb
f ccc

f is probably shorter than any key sequence that can pull up a previous command to edit :) Plus, the portion you want to change is now the final argument, making it simpler to edit if you do use up-arrow to bring up the last command.

CodePudding user response：

Thanks for the feedback everyone! I found another solution as well.

xargs -I {} grep {} file.txt <<< aaa

CodePudding user response：

Bash also supports substitution of the last command via ^old^new^:

$ grep shy lyrics.txt
You're too shy to say it
You're too shy to say it
$ ^shy^cry^
grep cry lyrics.txt
Never gonna make you cry
Never gonna make you cry
...