Bash read array from an external file

后端 未结 3 781
长情又很酷
长情又很酷 2020-12-04 02:32

I have setup a Bash menu script that also requires user input. These inputs are wrote (appended to) a text file named var.txt like so:

input[0]=\'192.0.0.1\'         


        
相关标签:
3条回答
  • 2020-12-04 02:52

    Use bash's readarray statement. (It's the only way I can find to put spaces in array elements dynamically.) You'll need your var.txt file to simply contain the elements of the array, one on each line, not contain assignment statements.

    readarray -t input < var.txt
    

    For more info, try help readarray (which will then tell you to try help mapfile).

    Here's my test for it:

    echo -e "a\nb c\nd" > var.txt
    readarray input < var.txt 
    for item in "${input[@]}"; do echo $item; done
    

    prints:

    a
    b c
    d
    

    Note: doing cat var.txt | readarray -t input doesn't work. I think it's because the input variable is scoped out of reach.

    0 讨论(0)
  • 2020-12-04 03:10

    You can encapsulate your variable extraction in a function and take advantage of the fact that declare creates local variables when used inside a function. This technique reads the file each time the function is called.

    readvar () {
        # call like this: readvar filename variable
        while read -r line
        do
            # you could do some validation here
            declare "$line"
        done < "$1"
        echo ${!2}
    }
    

    Given a file called "data" containing:

    input[0]='192.0.0.1'
    input[1]='username'
    input[2]='example.com'
    input[3]='/home/newuser'
    foo=bar
    bar=baz
    

    You could do:

    $ a=$(readvar data input[1])
    $ echo "$a"
    username
    $ readvar data foo
    bar
    

    This will read an array and rename it:

    readarray () {
        # call like this: readarray filename arrayname newname
        # newname may be omitted and will default to the existing name
        while read -r line
        do
            declare "$line"
        done < "$1"
        local d=$(declare -p $2)
        echo ${d/#declare -a $2/declare -a ${3:-$2}};
    }
    

    Examples:

    $ eval $(readarray data input output)
    $ echo ${output[2]}
    example.com
    $ echo ${output[0]}
    192.0.0.1
    $ eval $(readarray data input)
    $ echo ${input[3]}
    /home/newuser
    

    Doing it this way, you would only need to make one call to the function and the entire array would be available instead of having to make individual queries.

    0 讨论(0)
  • 2020-12-04 03:13

    If the whole var.txt file contains only Bash-compatible variable assignments as you indicated, you might just be able to source it, to make those variables available in a new Bash script:

    source var.txt
    
    useradd ${input[1]}
    

    This, however, will overwrite any existing variable with the same name. Command substitution can be used to avoid this, by selecting specific variables:

    input[1]="$(grep '^input\[1\]=' var.txt | sed "s|[^=]*='\(.*\)'|\1|")"
    

    It allows for renaming variables, although you will have to do this for each variable of interest. It essentially extracts the value of the variable from the var.txt file and assigns it to a new variable. See the grep manual page and the sed info page for more information on their use.

    Process substitution may allow for simpler expressions:

    source <(grep '^input\[[0-9]*\]=' var.txt)
    
    useradd ${input[1]}
    

    This would allow you to import only definitions of interest, although you have to watch for unwanted variable overwrites.

    0 讨论(0)
提交回复
热议问题