sorting

mohamad · 06-22-2010, 05:45 AM

hello
i wanna sort the 3rd column in a table numerically ( no actual borders...only tabs seperating the columns)
it should be something like this but i cant get it right

Sort -u +1 -3 results

the file is called results

jcomeau_ictx · 06-22-2010, 06:01 AM

Quote:

Originally Posted by mohamad

hello
i wanna sort the 3rd column in a table numerically ( no actual borders...only tabs seperating the columns)
it should be something like this but i cant get it right

Sort -u +1 -3 results

the file is called results

sort -n -k 3 results

sumeet inani · 06-22-2010, 06:36 AM

thanks jcomeau_ictx
i learnt that if we have a file containing table such that columns are separated by tabs then

Code:

$sort -n -k 3 input
means
-n numeric sort
-k 3 the the sorting column is third

I think only one thing should be kept in mind that elements of column before sort column must not contain space or tab as that is what whitespace is & is used in counting.

colucix · 06-22-2010, 07:16 AM

Quote:

Originally Posted by sumeet inani

I think only one thing should be kept in mind that elements of column before sort column must not contain space or tab as that is what whitespace is & is used in counting.

The -t option could help in this case. Anyway, for a TAB delimited file it's necessary some trick, since sort -t does not accept the \t notation. For example:

Code:

sort -t $'\t' -k3n file

or by typing a literal TAB using the Ctrl-V + TAB sequence to skip the shell's TAB completion:

Code:

sort -t "<Ctrl-V><TAB>" -k3n file

sumeet inani · 06-22-2010, 10:53 PM

Thanks colucix
Thus we can make table whose columns are separated by tabs provided the content of cell (atleast those which occur before sort column in the row) does not contain tab only spaces.

Why do you use -k3n ?

From what I understand in man page
-k startField,endField
When we don't mention end field then if content of sort column for two rows match then next column decides & so on...

pixellany · 06-22-2010, 11:08 PM

I have a hunch that "-k3n" is equivalent to "-k3 -n"

In general you can combine these "switches"--eg "ls -al" is the same as "ls -a -l" (And of course the ubiquitos "tar -xzvf"

Man pages plus experiments is the way to get your head around some of this.......

colucix · 06-23-2010, 12:45 AM

@sumeet inani: yes -k3n is a sort of combination of two options - as pixellany pointed out. However the start position and the end position can be more detailed, as explained in the man page when it talk about POS:

Code:

-k, --key=POS1[,POS2]
       start a key at POS1, end it at POS2 (origin 1)
<...>
POS is F[.C][OPTS], where F is the field number and C the character position  in  the  field;
both are origin 1.  If neither -t nor -b is in effect, characters in a field are counted from
the beginning of the preceding whitespace.   OPTS  is  one  or  more  single-letter  ordering
options,  which  override  global ordering options for that key.  If no key is given, use the
entire line as the key.

In -k3n we have F = 3, C = undef and OPTS = n.

sumeet inani · 06-23-2010, 06:46 AM

Understanding -k3n needs word by word reading of man page.
seems -k3n is similiar to -k 3 -n
Later is easier (though --key=3 --numeric-sort would be easiest but tedious) to understand & former fast to type.
Advanced users can write in cryptic ways.
I like simplicity.
Everyone has his way .