Deleting a subdirectory in the path of a URL via sed or perl #

Delete the divers subdirs which appear between "/blog/" and the filenames in this example:

$ cat index.rss
...
    <link>https://tinyapps.org/blog</link>
...
      <link>https://tinyapps.org/blog/misc/202009080700_usbimager.html</link>
...
      <link>https://tinyapps.org/blog/mac/202009070700_2011_mac_mini_snow_leopard.html</link>
...
      <link>https://tinyapps.org/blog/windows/202008200700_vostro_2420_lines.html</link>
...

with GNU sed:

$ sed -i -E 's (tinyapps.org/blog)/[a-z]+ \1 ' index.rss

-i = edit file in place
-E = use Extended regular expressions in the script
s = substitute the replacement string for the first instance of the regular expression in the pattern space. (Where are the forward slashes ("/") you ask? Replaced with whitespace! since sed accepts any single-byte character as a delimiter.)
(tinyapps.org/blog) = capture group (or, as the GNU sed documentation has it, subexpression)
/[a-z]+ = a forward slash followed by one or more lower case letters
\1 = back-reference referring to the subexpression above

or Perl:

$ perl -pi -e 's|(tinyapps.org/blog)/[a-z]+|\1|' index.rss

-p = print lines (more precise explanation)
-i = edit in-place
-e = enter one line of program
Unlike sed, perl does not allow whitespace as a delimiter ("you can use any pair of non-whitespace (ASCII) characters as delimiters"), so vertical bars were substituted.

/nix | Sep 13, 2020

Subscribe or visit the archives.