Splitting files in Unix using values in a file [closed]

Question

I have a txt file which has multiple XML records. I have to separate it using vendorname. I modified the commands as per my requirement. However, it's giving me 2 error messages:

awk: sf.awk:1: /<hdr:vendorname xmlns:hdr=\"http:\//www.mycompany.com\/header\/v1\">[A-Z]+<\/hdr:vendorname>/{
awk: sf.awk:1:                                         ^ syntax error
awk: sf.awk:1: /<hdr:vendorname xmlns:hdr=\"http:\//www.mycompany.com\/header\/v1\">[A-Z]+<\/hdr:vendorname>/{
awk: sf.awk:1:                                                               ^ backslash not last character on line

if I add a \ in front of ., then it gives me another error message

awk: sf.awk:1: /<hdr:vendorname xmlns:hdr=\"http:\//dwh\.www.mycompany.com\/header\/v1\">[A-Z]+<\/hdr:vendorname>/{
awk: sf.awk:1:                                         ^ backslash not last character on line

Script

/<hdr:vendorname xmlns:hdr=\"http:\//www.mycompany.com\/header\/v1\">[A-Z]+<\/hdr:vendorname>/{
    split($0, a, "hdr:vendorname xmlns:hdr=\"http:\//www.mycompany.com\/header\/v1">|<\/hdr:vendorname")
    if (out["file_"a[2]".txt"] == "") {
      out["file_"a[2]".txt"] = $0
    }
    else {
      out["file_"a[2]".txt"]=out["file_"a[2]".txt"]"\n"$0
    }
  }

END {
    for (fic in out) {
      printf out[fic] > fic
    }
  }

Answer 1

You need to backslash-escape both slashes in http://, not just the first one.

Thus, the pattern match should look like,

/<hdr:vendorname xmlns:hdr=\"http:\/\/www.mycompany.com\/header\/v1\">[A-Z]+<\/hdr:vendorname>/

Additionally, in your call to split(),

1. you shouldn't backslash-escape the forward slashes in the string;
2. you should backslash-escape the double-quote (") after v1.

So the string in split() should look like,

"hdr:vendorname xmlns:hdr=\"http://www.mycompany.com/header/v1\">|</hdr:vendorname"