Can field separator in awk encompass multiple characters?

Question

Can I use a field separator consisting of multiple characters? Like I want to separate words which contain quotes and commas between them viz.

\"School\",\"College\"

Answer 1

Try

awk 'BEGIN{FS="[|,:]"}{print $1}' youFile

Answer 2

Yes, you can use multiple characters for the -F argument because that value can be a regular expression. For example you can do things like:

echo "hello:::my:::friend" | gawk -F':::' '{print $3}'

which will return friend.

The support for regexp as the argument to -F is true for nawk and gawk (GNU awk), the original awk does not support it. On Solaris this distinction is important, on Linux it is not important because awk is effectively a link to gawk. I would therefore say it is best practice to invoke awk as gawk because then it will work across platforms.

Answer 3

yes, FS could be multi-characters. see the below test with your example:

kent$  echo '"School","College","City"'|awk -F'","|^"|"$' '{for(i=1;i<=NF;i++){if($i)print $i}}'
School
College
City

Answer 4

With GNU awk 4 you can easily parse even *CSV*s with embedded separators and quotes:

% cat infile 
"School",College: "My College","City, I"

% awk '{    
  for (i = 0; ++i <= NF;)
    print i, substr($i, 1, 1) == "\042" ?
      substr($i, 2, length($i) - 2) : $i
  }' FPAT='([^,]+)|(\"[^\"]+\")' infile  
1 School
2 College: "My College"
3 City, I

Answer 5

What's being talked around here is that the Field Separator isn't just limited to being multiple characters but can actually be a full-blown regex.

To wit: This strips out the header and surrounding tags from an XML fragment. Note that tags are well-formed, but different.

bash-3.2$ more xml_example 
<?xml version="1.0" encoding="UTF-8"?>
<urlset
xmlns="http://www.google.com/schemas/sitemap/0.84"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.google.com/schemas/sitemap/0.84
                  http://www.google.com/schemas/sitemap/0.84/sitemap.xsd">
<url>
<loc>http://www.foo.com/about.html</loc>
<lastmod>2006-05-15T13:43:37Z</lastmod>
<priority>0.5000</priority>
</url>
<url>
<loc>http://www.foo.com/articles/articles.html</loc>
<lastmod>2006-06-20T23:03:36Z</lastmod>
<priority>0.5000</priority>
</url>

Now we apply the awk script to print out the middle field, using a regex as the field separator:

bash-3.2$ awk -F"<(/?)[a-z]+>" '{print $2}' <xml_example




http://www.foo.com/about.html
2006-05-15T13:43:37Z
0.5000


http://www.foo.com/articles/articles.html
2006-06-20T23:03:36Z
0.5000

bash-3.2$

The blank lines are from where a tag was the only thing on that line, so there is no $2 to print. This is actually really powerful because it means that you can not only use fixed patterns with multiple characters but the full power of regular expressions as well in your field separator.