每当我找到相同的String时，如何浏览文件并设置数组？ 我正在使用bash代码进行编程

Question

I am a beginner in bash code. I am trying to set an array that contains different values that can be found in a file. The values I would like to add to the array are Strings that follow specific words. For example; Lets say the file is

Subcase 201 
Node 101 
"Much more text here" 
Subcase 202 
Node 501
"Again more text here"
Subcase 301 Node 804
.......

I would like to have the following output:

Subcase="201" "202" "301"; Node="101" "501" "804". 

Can someone help me please? I am not sure if I need to use a "for loop". I hope I were clear enough

Answer 1

What you are looking for is the read builtin. You can use read in a while loop. The following example reads from stdin (pipe the file to the script) but you can use read -r -u <fd> var vals to select from a different stream if you need.

#!/bin/bash
while read -r var vals
do
   eval P_${var}'+=($vals)' 2>/dev/null
done

for var in ${!P_@}
do
  eval echo '"${var#P_} = ${'$var'[@]}"'
done

The assignment line ignores stderr because var may contain invalid chars for a bash variable name.

Answer 2

Assuming that neither "Subcase" or "Node" appear twice on a given line of the original file, and that they are always spelled with the same case (first letter capitalized), the following code would achieve what you want:

#!/bin/bash

#assuming that a specific targeted string only appears once on a given line, and it is always written with same case (first letter capitalized). Cleaning data and storing all values on temporary files
grep -e 'Subcase' /folder/file > /tmp/subcase-holder && sed -i -e 's/^.*Subcase//' /tmp/subcase-holder && sed -i -e "s/^[ \t]*/\"/" /tmp/subcase-holder && sed -i -e 's/[[:blank:]]*$//' /tmp/subcase-holder && sed -i -e 's/\s.*$//' /tmp/subcase-holder && sed -i -e 's/$/\"/' /tmp/subcase-holder
grep -e 'Node' /folder/file > /tmp/node-holder && sed -i -e 's/^.*Node//' /tmp/node-holder && sed -i -e "s/^[ \t]*/\"/" /tmp/node-holder && sed -i -e 's/[[:blank:]]*$//' /tmp/node-holder && sed -i -e 's/\s.*$//' /tmp/node-holder && sed -i -e 's/$/\"/' /tmp/node-holder

#creating output file
cat /tmp/subcase-holder | tr '\n'  ' ' > /tmp/subcase-holder2 && sed -e 's/.$/; Node\=/' /tmp/subcase-holder2 > /tmp/partial-answer && sed -i -e 's/^/Subcase\=/' /tmp/partial-answer
cat /tmp/node-holder | tr '\n'  ' ' > /tmp/node-holder2 && sed -e 's/.$/\./' /tmp/node-holder2 > /tmp/partial-answer2
cat /tmp/partial-answer /tmp/partial-answer2 > /tmp/answer && echo >> /tmp/answer

exit 0

The result is stored on the file /tmp/answer. By the way, the code does not use arrays, but it delivers the output you want.