change the position of a line in a file using sed

Question

I would like to know how to change the position of a line in a file (preferably using sed). For example, consider the file that contains

goal identifier statement  
let statement 1  
let statement 2  
forall statement  
other statements

I would like to be able to do this

goal identifier statement  
forall statement  
let statement 1  
let statement 2  
other statements

where I change the position of the forall line and bring it after the goal line. forall and goal are regexps that can be used to identify the lines.

This is an interesting problem. To make it a good question on StackOverflow, please add the code you've tried to use to achieve a solution. Check out the [MCVE](http://stackoverflow.com/help/mcve) and [how-to-ask](http://stackoverflow.com/help/how-to-ask) pages for some guidelines on what makes a good question. — ghoti, Jan 28 '17 at 19:50

Jose Ricardo Bustos M. · Accepted Answer · 2017-01-28T19:28:54.537

6

you can try, for move line 4 to line 2, I want to move line A to line B, where A>B

sed -n '2{h; :a; n; 4{p;x;bb}; H; ba}; :b; p' file

or A<B

sed -n '2{h; d}; 4{p; x;}; p' file

you get, in first case: move line 4 to line 2

goal identifier statement
forall statement
let statement 1
let statement 2
other statements

you get, in second case: move line 2 to line 4

goal identifier statement  
let statement 2  
forall statement  
let statement 1  
other statements

Explanation

sed -n '              #silent option ON
    2{                #if is line 2
        h             #Replace the contents of the hold space with the contents of the pattern space
        :a            #label "a"
        n             #fetch the next line
        4{            #if is line 4
            p         #print line 4
            x         #Exchange the contents of the hold and pattern spaces
            bb        #goto "b"
        }
        H             #appends line from the pattern space to the hold space, with a newline before it.
        ba            #goto "a"
    }
    :b                #Label "b"
    p                 #print
' file

EDIT

If You want use regex for identify the lines, you can modify first command

sed -n '/goal/{p;n;h;:a;n;/forall/{p;x;bb};H;ba};:b;p' file

edited Jan 28 '17 at 19:28

answered Jan 28 '17 at 18:21

Jose Ricardo Bustos M.

8,016
6
40
62

1

I think they want to move lines based on regexp match not position? – 123 Jan 28 '17 at 18:55
@123 It's true, I misunderstood, I'm going to modify – Jose Ricardo Bustos M. Jan 28 '17 at 18:58
2

Nice could use a single branch though using something like `sed -n '/goal/{p;:1;n;/forall/!{H;b1};p;x;D};p'` – 123 Jan 28 '17 at 19:17
@Jose The regex (2nd command) works like a charm. Could you add the explanation for the regex command as well ? – Saurav Kumar Ghosh Jan 29 '17 at 02:57
@123 Works as well. Kindly explain the 'single branch' concept though. – Saurav Kumar Ghosh Jan 29 '17 at 02:59

score 1 · Answer 2 · answered Jan 28 '17 at 19:20

$ cat r.awk
BEGIN {
    forall_re = "^forall" # examples of regexps
    goal_re   = "^goal"
}

function tag(l) { # tag a line
    if      (l ~ goal_re  ) return "goal"
    else if (l ~ forall_re) return "forall"
    else                    return "rest"
}

{ # store entire file in array; give a tag to every line
    lines[NR] = $0 
    tags[NR]  = tag($0)
}

function swap0(a, i, j,   tmp) {
    tmp = a[i]; a[i] = a[j]; a[j] = tmp
}

function swap(i, j) {
    swap0(lines, i, j); swap0(tags, i, j)
}

function rise(i) {
    # TODO: add error check
    while (i - 1 > 0 && tags[i - 1] != "goal") {
        swap(i, i - 1); i--
    }
}

function process(    i) {
    for (i = 1; i <= NR; i++)
        if (tags[i] == "forall") rise(i)
}

function dump(    i) { # print the array
    for (i = 1; i <= NR; i++)
        print lines[i]
}

END {
    process()
    dump()
}

An example of input file

$ cat r.txt
goal identifier statement  
let statement 1  
let statement 2  
forall statement A  
other statements

goal identifier statement  
let statement 1  
let statement 2  
forall statement B
other statements

Usage:

$ awk -f r.awk r.txt
goal identifier statement  
forall statement A  
let statement 1  
let statement 2  
other statements

goal identifier statement  
forall statement B
let statement 1  
let statement 2  
other statements

score 0 · Answer 3 · answered Jan 30 '17 at 06:36

0

sed is for simple substitutions on individual lines, that is all. For anything else you should use awk for every desirable attribute of software (clarity, simplicity, portability, etc.:

$ awk 'NR==FNR{if (/forall/) {f=FNR; v=$0} next} FNR!=f; /goal/{print v} ' file file
goal identifier statement
forall statement
let statement 1
let statement 2
other statements

answered Jan 30 '17 at 06:36

Ed Morton

188,023
17
78
185

Would be helpful with an explanation of each awk item... – not2qubit Nov 24 '20 at 21:31
@not2qubit no, it's all trivial and obvious after a quick read of the awk manual. – Ed Morton Nov 24 '20 at 21:33

mug896 · Answer 4 · 2017-02-01T08:12:15.857

0

sed -r '/goal/{                 # if match "goal" line
    :X                          # this is a lable for branch command
    N                           # append next line
    /forall[^\n]*$/{            # if match "forall" line move to "goal" line below
        s#^([^\n]*)(.*)(\n[^\n]*)$#\1\3\2# 
        b                       # after move finished branch to end
    } 
    bX                          # branch to :X for appending next line
}' file
goal identifier statement  
forall statement  
let statement 1  
let statement 2  
other statements

edited Feb 01 '17 at 08:12

answered Jan 31 '17 at 19:38

mug896

1,777
1
19
17

Please consider adding a brief explanation of *how* the above solves the issue. – Leigh Feb 01 '17 at 01:51

score 0 · Answer 5 · answered Nov 24 '18 at 16:25

A less terrible way using vim to find a line in $FILENAME using regex $REGEX_EXPRESSION and move that line to $LINE_NUMBER:

vim -c "g:$REGEX_EXPRESSION:m$LINE_NUMBER" -cwq "$FILENAME"

explanation: -c is command in vim, so it goes to the first line that matches that regex and then moves it to the line number specified, and then does the command wq (or write and quit).

change the position of a line in a file using sed

5 Answers5

Linked