UNIX for Dummies Questions & Answers

Hi, I am trying to sort a file first by the string column, then by the number column.

file:

Code:

xyz1    2       
xyzX    4        
xyz2    1       
xyz13   3       
xyz11   5  
xyz13  10     
xyz1    1        
xyz10   1        
xyz4    2

result should be

Code:

xyz1    1    
xyz1    2  
xyz2    1  
xyz4    2 
xyz10   1  
xyz11   5    
xyz13   3  
xyz13   10 
xyzX    4

I have tried many sort commands and failed. Please help!

---------- Post updated at 02:45 PM ---------- Previous update was at 12:32 AM ----------

please help!

try this:

Code:

> awk 'BEGIN {OFS="|"} {print $1,$2}' filename | sort -n | awk -F'|' 'BEGIN {OFS=" "}{print $1,$2}' > newfile

Thanks. I think I have figured it out. The file is very large, so I need to use the unix sort.(not sure about the awk thing) I used sort -t ' ' -k

You have to specify a numerical sort for the second column otherwise it will get sorted alphabetically. This should do it:

Code:

sort -k1,1 -k2,2n infile

With my sort you have to use a comma in the key field specification or it does not work right. Also, if you use -t' ' it seems to use single space separation instead of multispace separation and also produce incorrect result. Without -t it produces:

Code:

xyz1    1
xyz1    2
xyz10   1
xyz11   5
xyz13   3
xyz13   10
xyz2    1
xyz4    2
xyzX    4

But you required a numerical sort on the second part of the first key. For that the first key needs to be split into an alphabetical part and a numerical part during the sort phase. A numerical sort places X at the top though, not at the bottom.

Code:

sed 's/\(...\)/\1 /' infile|sort -k1,1 -k2,2n -k3,3n|sed 's/ //'

produces:

Code:

xyzX    4
xyz1    1
xyz1    2
xyz2    1
xyz4    2
xyz10   1
xyz11   5
xyz13   3
xyz13   10