CODEWITHSUNDEEP
sortingunixcommand-line-interface
oliver
oliver

Reputation: 2833

unix 'sort' command for inline characters

I have a .txt file of pumpkinsizes that I'm trying to sort by size of pumpkin:

name |size
==========
Joe  |5
Mary |10
Bill |2
Jill |1
Adam |20
Mar  |5
Roe  |10
Mir  |3
Foo  |9
Bar  |12
Baz  |0

Currently I'm having great difficulty in getting sort to work properly. Can anyone help me sort my list by pumpkin size without modifying the list structure?

Upvotes: 1

Views: 126

Answers (4)

catssay
catssay

Reputation: 106

The key point is the option -k of sort. You can use man sort to see how it works. The solution for your problem follows:

sed -n '3,$p' YOUR_FILENAME| sort -hrt '|' -k 2

Upvotes: 3

glenn jackman
glenn jackman

Reputation: 246774

You can do this in the shell:

{ read; echo "$REPLY"; read; echo "$REPLY"; sort -t'|' -k2n; } < pumpkins.txt

That reads and prints the first 2 header lines, then sorts the rest.

Upvotes: 1

sameerkn
sameerkn

Reputation: 2259

You can simply remove the 

name |size
==========

by using sed command. Then whatever is left can be sorted using sort command.

sed '1,2d' txt | sort -t "|" -k 2 -n

Here, sed '1,2d' will remove the first 2 lines.

Then sort will tokenize the data on character '|' using option -t.

Since you want to sort based on size which happens to be second token, so the token "size" can be specified by -k 2 option of sort.

Finally, considering "size" as number, so this can be specified by option -n of sort.

Upvotes: 1

Jens
Jens

Reputation: 72639

The table headings need special consideration, since "sorting" them will move them to some random line. So we use a two step process: a) output the table headings. b) sort the rest numerically (-n), reverse order (-r), with field separator | (-t), starting at field 2 (-k)

$ awk 'NR<=2' in; awk 'NR>2' in | sort -t '|' -nr -k 2
name |size
==========
Adam |20
Bar  |12
Roe  |10
Mary |10
Foo  |9
Mar  |5
Joe  |5
Mir  |3
Bill |2
Jill |1
Baz  |0

Upvotes: 3

Related Questions