grep based on condition then sed-CodePudding

I have a file and I want to write a command that deletes the block based on the word. If I want to delete [email protected] the deletion start from the word user to } character.

file:

user "sara" {
  first_name           = "Sara"
  login                = "[email protected]"

}

user "john_smith" {
  first_name           = "John"
  login                = "[email protected]"
  last_name            = "Smith"

}

I am a beginner at sed and grep commands

I would appreciate it if someone help!

CodePudding user response：

It might be a little difficult to get the results you're looking for using just grep and sed. If you're able to use perl, though, this should work:

# cat test.txt
user "sara" {
  first_name           = "Sara"
  login                = "[email protected]"

}

user "john_smith" {
  first_name           = "John"
  login                = "[email protected]"
  last_name            = "Smith"

}

user "sara" {
  first_name           = "Sara"
  login                = "[email protected]"

}

user "john_doe" {
  first_name           = "John"
  login                = "[email protected]"
  last_name            = "Doe"
}
# perl -pe 'BEGIN {$/ = "}"} undef $_ if /sara\@example.com/' test.txt


user "john_smith" {
  first_name           = "John"
  login                = "[email protected]"
  last_name            = "Smith"

}

user "john_doe" {
  first_name           = "John"
  login                = "[email protected]"
  last_name            = "Doe"
}

Commented code below:

# Run with the -p flag to, by default, print every line in the file being read
BEGIN {
    # Set the INPUT_RECORD_SEPARATOR variable to a closing curly brace
    # This causes Perl to treat each block as a separate "line"
    $/ = "}";
}
# If the current block contains "[email protected]", don't print anything.
# Note that the ampersand (@) character has to be escaped with a backslash as seen below
undef $_ if /sara\@example.com/;

CodePudding user response：

This might work for you (GNU sed):

sed '/user/{:a;/}/!{N;ba};/sara@example\.com/d}' file

Gather up lines starting with a line that contains user and ending when another (or the same) line contains }.

If the collection contains the string [email protected], then delete those lines.