如何使用字符串和随机生成的数字替换文件中的标头值？

Question

I Have files in a directory that look like:

aaa   bbbbb   cccc  value1=xxxxxvalue2=yyyy
body body body
body body body 
body body body 
aaa   bbbbb   cccc  value1=xxxx

How do I replace everything between "value1=" and "value2=" in the header with a string (sevens.eighteen. below) and a randomly generated number of length 20 (11234567890123456789 below)?

ex: aaa bbbbb cccc value1=sevens.eighteen.11234567890123456789=yyyy

I have tried variations of the below, but I am not getting what I want.

sed -i '1 s/^.*value=/yoursubstitution value=/; $ s/^.*value=/yoursubstitution value=/'

Here are results I want to see in all of the files in that directory:

aaa   bbbbb   cccc  value1=sevens.eighteen.11234567890123456789value2=yyyy
body body body
body body body 
body body body 
aaa   bbbbb   cccc  value1=xxxx

Answer 1

With one file.

# create a 20 digit random number with bash
for ((i=0;i<20;i++)); do num="${num}${RANDOM: -1:1}"; done

# replace string with GNU sed
sed -E '1 s/(value1=).*(value2=.*)/\1sevens.eighteen.'"$num"'\2/' file

Output:

aaa   bbbbb   cccc  value1=sevens.eighteen.39489360437938728734value2=yyyy
body body body
body body body 
body body body 
aaa   bbbbb   cccc  value1=xxxx

Answer 2

There are different ways to create a random digit:

RAND_NR="$(tr -cd "[:digit:]" < /dev/urandom | head -c 20)"

Then you can replace the value using sed :

sed "s/value1=.*value2=/value1=sevens.eighteen.${RAND_NR}value2=/" file

Double quotes in sed to expand the variable RAND_NR.

The output:

aaa   bbbbb   cccc  value1=sevens.eighteen.76744539817863807290value2=yyyy
body body body
body body body 
body body body 
aaa   bbbbb   cccc  value1=xxxx

Answer 3

With GNU awk for the 3rd arg to match() and -i inplace for "inplace" editing you just need this one command:

$ cat tst.awk
BEGIN { srand(seed) }
(FNR==1) && match($0,/(.*value1=).*(value2=.*)/,a) {
    $0 = a[1] "sevens.eighteen." int(10000000000000000000 + rand()*90000000000000000000) a[2]
}
{ print }

$ awk -i inplace -v seed="$RANDOM" -f tst.awk *

For example:

$ awk -v seed="$RANDOM" -f tst.awk file
aaa   bbbbb   cccc  value1=sevens.eighteen.77242153065798582272value2=yyyy
body body body
body body body
body body body
aaa   bbbbb   cccc  value1=xxxx

$ awk -v seed="$RANDOM" -f tst.awk file
aaa   bbbbb   cccc  value1=sevens.eighteen.59551444957943291904value2=yyyy
body body body
body body body
body body body
aaa   bbbbb   cccc  value1=xxxx

Note that the above will only operate on the first line of each input file so even if the regexp matched on a different line it wouldn't be changed.

I see you asked another identical question other than you want to do it for the first and last lines. Just move the code that changes $0 to a function since you'll need to call it twice now and save/change the previous line rather than the current line so you have the last line to work on in ENDFILE (also a GNU awk extension):

$ cat file
aaa   bbbbb   cccc  value1=xxxxxvalue2=yyyy
body body body
body body body
body body body
aaa   bbbbb   cccc  value1=xxxx,value2=yyyy

$ cat tst.awk
BEGIN { srand(seed) }
NR==1  { fmt() }
NR > 1 { print prev }
{ prev = $0 }
ENDFILE { fmt(); print }

function fmt() {
    if ( match($0,/(.*value1=).*(value2=.*)/,a) ) {
        $0 = a[1] "sevens.eighteen." int(10000000000000000000 + rand()*90000000000000000000) a[2]
    }
}

$ awk -v seed="$RANDOM" -f tst.awk file
aaa   bbbbb   cccc  value1=sevens.eighteen.55152151176931606528value2=yyyy
body body body
body body body
body body body
aaa   bbbbb   cccc  value1=sevens.eighteen.98796005800472494080value2=yyyy