0

I have the following CSV file as shown below with 4 columns, and I am taking input of filename from bash terminal.

Sile,Barhams,Female,[email protected]
Shae,Scotti,Male,[email protected]
Shae,Bonham,Female,[email protected]
Shay,Ert,Female,[email protected]
Simeon,Dearle,Male,[email protected]

Desired output is:

Shae,Male
Shae,Female
Shay,Female
Sile,Female
Simeon,Male

Now I tried following code snippets and got the result as shown :

cat $file | cut -d, -f1,3 | sort -k 1,1 -s

Shae,Female
Shae,Male
Shay,Female
Sile,Female
Simeon,Male

cat $file  | cut -d, -f1,3 | sort -k 1,1 -s -n

Sile,Female
Shae,Male
Shae,Female
Shay,Female
Simeon,Male

Can somebody explain my both snippets and how I can get desired one please, according to me for first code I am sorting according to first column and in case of tie I have specified stable sort to get the desired output but IDK what is the issue and in second code I wonder how "Sile" appears on the top.

Improve this question
4
  • If supported by your sort implementation, the --debug option will likely be of help. Commented Apr 9, 2021 at 11:43
  • Please tell me what changes i need to do for getting desired output... I need help as ai tried all sort of possible combinations of flags of sort Commented Apr 9, 2021 at 12:15
  • thnx @fra-san infact -t, worked but I would request you that how it worked ? and moreover explain my pipeline of both code that I stated above as a new answer please it will improve my understanding. Commented Apr 9, 2021 at 13:04
  • ah, indeed, I missed the thing with the two Shae's switching order the first time. Sorry. Commented Apr 9, 2021 at 13:40

2 Answers 2

Reset to default
2 
cat $file | cut -d, -f1,3 | sort -k 1,1 -s

Shae,Female
Shae,Male
Shay,Female
Sile,Female
Simeon,Male

You get that because the fields -k refers to are the blank separated fields by default and Shae,Female sorts before Shae,Male.

You need the -t option to tell it to use the comma as a separator instead:

-t, --field-separator=SEP
use SEP instead of non-blank to blank transition

$ cat "$file"  | cut -d, -f1,3 | sort -t, -k 1,1 -s
Shae,Male
Shae,Female
Shay,Female
Sile,Female
Simeon,Male

cat $file  | cut -d, -f1,3 | sort -k 1,1 -s -n

Sile,Female
Shae,Male
Shae,Female
Shay,Female
Simeon,Male

Here, with -n, the sorting is based on the numerical value of the data. But you don't have a numeric strings here: all non-numeric strings will be treated as zeroes. Hence, effectively no primary sorting rule. With -s, the order will be left as-is. Without -s, sort will fall back to a textual sort to get some result.

Improve this answer
1

Strange. When I apply your pipeline to that data, I get

Sile
Shae
Shae
Shay
Simeon

Let's have a closer look at the pipeline:

cat filename | grep "^.\{0\}S" | cut -d, -f1,6 | sort -s -n -k1,1

First, simplify it to

grep "^S" filename | cut -d, -f1,6 | sort -s -n -k1,1

You don't need the cat (it's sometimes called "cat abuse"), and the regular expression asks for zero characters after the beginning of a line, followed by S. This is equivalent to a line that starts with S.

Why do you cut out fields 1 and 6 when there are only two fields? And I don't believe that the result keeps the second field. Something is not right here, but let's ignore the strange cut.

You perform a numerical sort. Since there is no number in the file and -s is specified, the order remains as it is. Remove the -n and/or the -s.

Improve this answer
3
  • In particular, it's the combination of numeric and stable that gives the result that nothing happens. Without -s, it would sort the lines with equal numerical value (i.e. all) based on the text order (well, my sort does). Commented Apr 9, 2021 at 10:26
  • I modified question please now do the needful and I would request to remove previous comments so to have have more clear answer. Commented Apr 9, 2021 at 11:11
  • The first snippet has the desired result. My answer explains why. I can't explain the cat, though. Commented Apr 9, 2021 at 13:20

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.