mjhjmtu

i am trying to delete differnt values from a json file. Its not pretty and jq return 1 compile error". So i searching for a solution within bash (awk/sed).

Here is my sample data from my json-file (similiar to docker-compose):

aaaa"MyPorts":"80/tcp":,"8080/tcp":,"9080/tcp":,"9090/tcp":,aaaa 
bbbb"PROTO 9080/tcp 9090/tcp"bbbbb
cccc"TEST PROTO 80 8080"ccccc

all entries can be found mutiple times or also within one line.

After using the replace it should look like:

aaaaaaaa 
bbbbbbbbb
cccc"TEST"ccccc

I tried sed and (awk -F["]) but could not get some useful results.
How can this be done ?

Update: i made the sample clearer..

There are 3 different types:...:

1: Delete from "MyPorts" to }},
2. Delete "Proto to "
3. Delete from " PROTO" till one character before ", only if left from Proto is a space 

Update 2: How can this be done if the sample data is also in one line, like this:

aaaa"MyPorts":"80/tcp":,"8080/tcp":,"9080/tcp":,"9090/tcp":,aaaabbbb"PROTO 9080/tcp 9090/tcp"bbbbbcccc"TEST PROTO 80 8080"ccccc

After using the replace it should look like:

aaaaaaaabbbbbbbbbcccc"TEST"ccccc

sed approach:

sed -E 's~("MyPorts.*},|"PROTO[^"]+"| +PROTO[^"]+)~~' file

The output:

aaaaaaaa 
bbbbbbbbb
cccc"TEST"ccccc

Bonus approach for one-line string input:

$ s='aaaa"MyPorts":"80/tcp":,"8080/tcp":,"9080/tcp":,"9090/tcp":,aaaabbbb"PROTO 9080/tcp 9090/tcp"bbbbbcccc"TES^CPROTO 80 8080"ccccc'

$ sed -En 's~("MyPorts.+},|"PROTO[^"]+"| +PROTO[^"]+)~~gp' <<<"$s"
aaaaaaaabbbbbbbbbcccc"TEST"ccccc