Clever Geek Handbook

Regex matches end of line

I'm looking for a BASH regex to pull the db commands from the below commands. However, the order of the arguments is not guaranteed. For some reason, I cannot get it to work completely.

What am I still

regex="--db (.*)($| --)"
[[ $@ =~ $regex ]]
DB_NAMES="${BASH_REMATCH[1]}"

# These are example lines
somecommand --db myDB --conf /var/home # should get "myDB"
somecommand --db myDB anotherDB manymoreDB --conf /home # should get "myDB anotherDB manymoreDB" 
somecommand --db myDB # should get "myDB"
somecommand --db myDB anotherDB # should get "myDB anotherDB"

Any suggestion for regex?

+3


source to share


4 answers


The problem is that it bash

uses a flavor regex

that doesn't include the unwanted repetition operators ( *?

, +?

). Since it *

is greedy, and there is no way to tell it not to be greedy, the first subexpression in parentheses ( (.*)

) matches everything up to the end of the line.

You can get around this if you know that the values ​​you want to capture do not contain a specific character and are replaced .

with a character class that excludes that character.

For example, if the values ​​after --db

do not contain a dash ( -

), you can use this one regex

:



regex='--db ([^-]*)($| --)'

It matches all the examples posted in the question.

+4


source


The following works:

regex="--db[[:space:]]([[:alnum:][:space:]]+)([[:space:]]--|$)"
[[ "$@" =~ $regex ]]


There were two questions:

  • Character classes such as [: space:] should be used to represent spaces
  • (.*)

    is greedy and will get to your last --

    literal. Since bash does not support unwanted matching, we have to match using [[:alnum:][:space:]]

    that will ensure we stop at the next one --

    .
+2


source


By default RegEx tries to get as many matches as possible, use a nonliving (lazy) quantifier. You can also set first --

so that the engine will use that first

--db[[:space:]](.*?)([[:space:]]--|$)

Demo

<h / "> If you don't want --

, you can use the group without capturing

--db[[:space:]](.*?)(?:[[:space:]]--|$)
                     ^^ Notice the ?:

Demo

+1


source


I think you want to match characters without spaces to catch the first group:

regex="--db (\S+)( --|$)"
0


source






More articles:


All Articles