Question

I’m trying to strip all the lines in an XML file that lie between the tags < KEYS > and < /KEYS >.

As a first pass at the problem I’ve gotten a regex that will match the first block of keys in the file but it doesn’t continue matching the other blocks in the file. I’ve tried adding “/g” to the regex and I’ve tried “-0777” to slurp the whole file at once and neither trick makes any difference. Below is the perl one liner:

perl -00 -ne 'print $1 if /(\s+\<KEYS\>\n\s+.*?\n\s+\<\/KEYS\>)/s' someFile.xml

and I get this output:

  <KEYS>
    <KEY name="cone_id" type="long" nativeType="number(17)"/>
    <KEY name="bar_id" type="long" nativeType="number(32)"/>
    <KEY name="foo_type" type="int" nativeType="number(3)"/>
  </KEYS>

As stated above there are a lot more blocks in the file (which is nearly five thousand lines long) but the perl code isn’t messing with any of the rest.

Any suggestions?

Answer 1

Your one-liner will do what you want by making two changes:

1. Change if to while.
2. Add the g option to your regex: /.../gs

Alternatively, it looks like the start and end tags of interest are on different lines, by themselves. If so, the flip-flop operator could be handy:

perl -ne 'print if m{<KEYS>} .. m{</KEYS>}' DATA_FILE