使用sed，在文本文件中标识右行，并将三个不同的部分写入以分隔变量

Question

I have a group of text files that each contain one instance of the following line somewhere, where the "(MR #" is the best way to identify the right line...

Smith, John A (MR # MR123456)

I'd like to get three sed statements that, when run against a given file, populate the following three variables: Plname, Pfname and MRnum. Using the example above, I'd like the variables to end up holding the following after running the sed statements.

Plname=Smith
Pfname=John
MRnum=MR123456

Despite many sed examples out there, I'm new to this, and currently struggling with the exact syntax. Thanks for your help.

Answer 1

Something like this:

$ cat t
bla-bla-bla
Smith, John A (MR # MR123456)
bla-bla-bla
$ Plname="$(sed -n '/(MR #/{s/^\([^,]\+\),.*/\1/p;q}' t)"
$ Pfname="$(sed -n '/(MR #/{s/^[^,]\+,[ ]\?\([^(]\+\).*/\1/p;q}' t)"
$ MRnum="$(sed -n '/(MR #/{s/^[^(]\+(MR # \([^)]\+\).*/\1/p;q}' t)"
$ printf "Plname = %s, Pfname = %s, MRnum = %s\n" "$Plname" "$Pfname" "$MRnum"
Plname = Smith, Pfname = John A , MRnum = MR123456

---

Little explanation:

/(MR #/{s/^\\([^,]\\+\\),.*/\\1/p;q} is a short form of

/(MR #/ #1 { s/^\\([^,]\\+\\),.*/\\1/; #2 p; #3 q #4 }

• sed will parse file line by line
• -n flag says that sed won't print each line (it does it by default), because we will do this manually if needed
• #1 ( // ) finds line which contains (MR #
• and if line matches then we do following actions:
  • #2 ( s/// ) replaces its content by regular expression
  • #3 ( p ) prints result
  • #4 ( q ) stops processing of file because we already find what we are looking for