Shell Programming and Scripting

Interesting.
How does this work?

Code:

$1=$1    #$1 is the first column

---------- Post updated at 05:40 PM ---------- Previous update was at 05:36 PM ----------

Just do it:

Code:

awk '{print $1}' file
awk '{print $1" "$2}' file
awk '{$1=$1;print}' file

$1=$1 is a typical awk idiom to force awk to rebuild $0. mostly used to apply a new OFS.
It removes all unnecessary double spaces.

That shouldn't be so surprising. They are just different methods of arriving at the same solution. Just as there are different routes to go from point A to point B.

The regex in the first awk solution used a character class with the repetition quantifier "+".

Code:

/^[ 	]+/

There's a blank space and a tab character in there which means that this regex searches for one or more occurrences, from the beginning, of either a space or tab character (or any combination of both). The sub function substitutes those with a zero-length string.
I think "\t" for the actual tab character should work; at least it does for gawk:

Code:

gawk '{sub(/^[ \t]+/,""); print}' file

The "+" repetition quantifier is an Extended Regular Expression (ERE). Gnu sed allows it, but the sed binaries in most Unix systems do not. So the other type of regex used in the sed solution was this -

Code:

/^ \{1,\}/

The bracket repetition operator is for finer control over repetition. "+" means one or more - there's no limit for "more". Whereas {m,n} means at least m at the most n repetitions. Further, {m,} means at least m repetitions - there's no upper limit here. So {1,} becomes equivalent to "+", which is why the sed solution works more or less the same. (It doesn't take care of tab characters though).

Perl is a different beast altogether. All the above concepts, BREs as well as EREs, are for POSIX regexes. Perl, on the other hand, started off by implementing Henry Spencer's regular expression library. It's regex syntax is richer, more consistent and more extensive than those of POSIX compliant regexes.

HTH,
tyler_durden

this is an very excellent resource, which will be very handy to you while working across several commands: http://www.greenend.org.uk/rjk/2002/06/regexp.html