Replacing multiline with sed

user_402866 0
Viewed 37

I want to replace

@Column("uuid", {
   primary: true,
   name: "id",
   default: () => "uuid_generate_v4()",
})

with

@PrimaryGeneratedColumn

in a file with the following content:

export class SomeClass {
  @Column("uuid", {
    primary: true,
    name: "id",
    default: () => "uuid_generate_v4()",
  })
  id: string;

  @Column("timestamp with time zone", {
    name: "created_at",
    default: () => "now()",
  })
  created_at: Date;

  @Column("timestamp with time zone", {
    name: "updated_at",
    default: () => "now()",
  })
  updated_at: Date;

  @Column("uuid", { name: "other_id", unique: true })
  other_id: string;
}

This is my current solution:

sed -E -i '' -e '/@Column\(\"uuid\", \{/,/\}\)?/c  @PrimaryGeneratedColumn\(\"uuid\"\)' fileName

It basically works, but it's too inaccurate and also replaces other lines that shouldn't be replaced. I tried to add uuid_generate_v4() but couldn't get it to work, probably because of the line breaks.

sed
2 Answers

This might work for you (GNU sed):

sed '/@Column("uuid"/{:a;N;/^\s*})$/M!ba;/"uuid_generate_v4()/s/.*/@PrimaryGeneratedColum/}' file

Gather up lines between those containing @Column("uuid" and }) and if those lines also contain uid_generate_v4(), replace all them by PrimaryGeneratedColum.

user_9674920

Using sed, you can anchor the end of the range to stop it matching with with the duplicate range.

$ sed -i '' -e '/@Column("uuid", {$/,/})/c\@PrimaryGeneratedColumn' input_file 
rt class SomeClass {
@PrimaryGeneratedColumn
  id: string;

  @Column("timestamp with time zone", {
    name: "created_at",
    default: () => "now()",
  })
  created_at: Date;

  @Column("timestamp with time zone", {
    name: "updated_at",
    default: () => "now()",
  })
  updated_at: Date;

  @Column("uuid", { name: "other_id", unique: true })
  other_id: string;
}
user_163721090
Related