Help in removing leading spaces in a field 2

[rufflocks]

I have a pipe-delimited file that has 1 or more leading spaces. Do you know how to remove these leading spaces, using AWK, while I'm proceesing the file?

ex: 987| 234| 2| 567

 

[grega]

Assuming you want to remove all spaces from the line, you can use the gsub function in nawk.

nawk -F"|" '{gsub(" ","",$0);print $1,$2,$3,$4}' filename

Greg.

 

[Krunek]

Hi!

You can omit $0 in gsub function. If 3rd parameter of the gsub is not supplied, it defaults to $0.

Please, try the same example as grega's but littlebit different:

awk 'BEGIN { RS = "|" } { gsub(" ","" print }' inputfile

Bye!

KP.

 

[rufflocks]

Thx grega. I do not want to delete all spaces, just the leading spaces. Some of the fields will have name (|first <space> last|) and address. So only leading spaces will be deleted.

 

[vgersh99]

for (i=1; i <= NF; i++)
nsub1=gsub(&quot;^[[:space:]]+|[[:space:]]+$&quot;, &quot;&quot;, $i);

 

[CaKiwi]

Another option is to use sed as follows:

sed &quot;s/\(|\) */\1/&quot; filename

If you also need to remove spaces at the beginning of each line, pipe the output into another sed like so:

sed &quot;s/\(|\) */\1/&quot; filename | sed &quot;s/^ *//&quot;

Hope this is helpful.

CaKiwi

 

[CaKiwi]

But if you insist on using awk, this should work.

awk '{gsub(&quot;\| *&quot;,&quot;|&quot;gsub(&quot;^ *&quot;,&quot;&quot;print}' filename

CaKiwi

 

[rufflocks]

CaKiwi,

sed &quot;s/\(|\) */\1/&quot; filename works for me, Thanks.

Could you please explain it in English. I know that it removes leading spaces that immediately follow a pipe, but what does each piece following sed&quot;s/ mean?

Thanks in advance.

 

[CaKiwi]

The \( \) pair save everything between them (in this case just the vertical bar) in register 1, which is replayed in the substitution by the \1. An asterisk following a character matches any number of that character so the blank asterisk matches any number of blanks.

Hope this helps.

CaKiwi

 

[Krunek]

Well,

I tried to solve this problem with original version of awk that is described in this document:

http://ortega.cs.ucdavis.edu/~byrav/Mozz/awk.doc;

gsub() function isn't avaible in original version of awk, so I must write a bunch of code:


BEGIN { FS = &quot;|&quot; }

{
# removing leading spaces
for (i = 1; i <= NF; i ++) { # any field of input line
n = split($i, ch, &quot; &quot
for (j = 1; j <= n; j ++) { # any word of the field
if (j == 1)
newFld = ch[j]
else
newFld = newFld &quot; &quot; ch[j]
}
if (newLn == &quot;&quot
newLn = newFld
else
newLn = newLn FS newFld
}

# Contens of the variable &quot;newLn&quot; is without leading spaces
# in the fields.
# Here is posible to process the contens
# of the variable &quot;newLn&quot; (instead &quot;print newLn&quot.
print newLn

newLn = &quot;&quot;
}

Bye!

KP.

 

[CaKiwi]

Although now that I look at it again it would have been easier to write:

sed &quot;s/| */|/&quot; filename

and to also remove leading blanks:

sed &quot;s/| */|/;s/^ //&quot; filename

CaKiwi