find a replace multiple strings in a matching code block using sed - linux

Question

Please any help on this using Linux "sed" will be appreciated.
I have the code block as below in a file. Looking to extract a block starting with "aa.bb.cc.config=" and end with "}". From the extracted block, find the strings ["bb11:11211"] ["ff5rt:11211"] and delete the matching lines from the file.
asdf={
ddd, \
fff, \
ggg \
}
ghjk
wwww
aa.bb.cc.config={"List": [ \
"aa11:222", \
"bb11:11211", \
"cc22ss:11211", \
"dd33ss:11211", \
"ff5rt:11211", \
"gg6hj:11211", \
], "dd": true, "ff": 0, "hh": 0, "jj": 10, "kk": 0 }
vsdfsdf
ewqewqe
OUTPUT:
asdf={
ddd, \
fff, \
ggg \
}
ghjk
wwww
aa.bb.cc.config={"List": [ \
"aa11:222", \
"cc22ss:11211", \
"dd33ss:11211", \
"gg6hj:11211", \
], "dd": true, "ff": 0, "hh": 0, "jj": 10, "kk": 0 }
vsdfsdf
ewqewqe
Thanks in Advance!!

Answer 1

sed approach:
sed '/^aa\.bb\.cc\.config=/,/}$/{ /\"\(bb11\|ff5rt\):11211\"/d; }' file
/^aa.bb.cc.config=/,/}$/ - address range, considering lines from /^aa.bb.cc.config=/ to /}$/
/\"\(bb11\|ff5rt\):11211\"/d; - delete lines matched the pattern within captured pattern space
The same using extended regex -E:
sed -E '/^aa\.bb\.cc\.config=/,/\}$/{ /"(bb11|ff5rt):11211"/d; }' file
The output:
asdf={
ddd, \
fff, \
ggg \
}
ghjk
wwww
aa.bb.cc.config={"List": [ \
"aa11:222", \
"cc22ss:11211", \
"dd33ss:11211", \
"gg6hj:11211", \
], "dd": true, "ff": 0, "hh": 0, "jj": 10, "kk": 0 }
vsdfsdf
ewqewqe