Remove String Captured in Regex Group - sed

Question

I have a file that contains the following content with tags/tag/volume_tags. I want to use sed command to remove stackId = AnyValue inside tags/tag only (not volume_tags).
Input:
tags = {
Name = var.stack_id
StackId = bar
}
tags =
{
Name = var.stack_id
StackId = bar
}
tag = {
StackId = foo
}
volume_tags = {
StackId = bar
}
Output:
tags = {
Name = var.stack_id
}
tags =
{
Name = var.stack_id
}
tag = {
}
volume_tags = {
StackId = bar
}
I try creating Regex that captures the tags/tag block and group StackId = AnyValue with (\s*StackId.\s*=.*) and aim to use sed command to replace that group with an empty string. How to do so using sed? or any other suggestion to achieve this?
^tag.*\s*=?\s*{(\s*.*\s*)?(\s*StackId.\s*=.*)\s*}

Answer 1

This might work for you (GNU sed):
sed '/^tags\?\>/{:a;n;/\<StackId\>/d;/^}$/!ba}' file
If a line begins with tag or tags fetch the following lines until one beginning/ending }.
If the collection of lines above contains the word StackId delete that line only.
N.B. Assumes only one such id per stanza.
cat <<\! | sed -f - file
/^tags\?\>/{
:a
n
/\<StackId\>/d
/^}$/!ba
}
!

Answer 2

sed is line-oriented. It can perform actions on ranges of lines:
sed -e '
/^[[:space:]]*tags\{0,1\}[[:space:]]*=[[:space:]]*{[[:space:]]*$/, /^[[:space:]]*}[[:space:]]*$/ {
/^[[:space:]]*StackId[[:space:]]*=/ d
}
'
\{0,1\} is equivalent to more usual ? in other regex versions
sed can't portably do case-insensitive matching but there are workarounds