I have some file like this :
abc 123 abc 789 bcd 456 acb 135 I would like to print first column of next line in current line.
Desired output:
abc 123 abc abc 789 bcd bcd 456 acb acb 135 I prefer to use awk.
Add a comment |
4 Answers
Memorise the previous line:
awk 'NR > 1 { print prev, $1 } { prev = $0 } END { print prev }' This processes the input as follows:
- if the current line is the second or greater, print the previous line (stored in
prev, see the next step) and the first field of the current line, separated by the output field separator (the space character by default); - in all cases, store the current line in the
prevvariable; - at the end of the file, print the previous line.
Alternative awk approach:
$ awk 'NR == 1{printf "%s", $0;next}{printf " %s\n%s", $1,$0}' input.txt abc 123 abc abc 789 bcd bcd 456 acb acb 135 The way this works is simple: first line is special case - we print it without new-line, and tell awk to go to next line without executing other code blocks. After that, NR == 1{printf "%s", $0;next} is skipped, but other parts do the job.
Remember that up to now we printed a formatted string without new line character. Thus , what is being done by printf " %s\n%s",$1,$0 now is first word is printed out ( and because there was no newline, it remains on the same line of output) , newline inserted, and then whole line itself ( but doesn't terminate with newline character). Thus next first word inserted will remain on the same line. Process continues on and on till we reach the end of file.
Possible improvement is to include END{print ""} block to insert final newline. In certain cases where resulting file is to be processed by other scripts it might be desirable.
While the user requested AWK specifically, same approach with printing formatted strings can be taken with other languages, for instance Python. Python alternative provided for those curious about how this can be implemented in other languages:
#!/usr/bin/env python from __future__ import print_function import sys old = None for index,line in enumerate(sys.stdin): if index == 0: print(line.strip(),end=" ") continue words = line.strip().split() print(words[0] + "\n" + line.strip(),end=" ") And usage like so:
$ ./append_first.py < input.txt abc 123 abc abc 789 bcd bcd 456 acb acb 135 Same idea about final newline applies here.
answered Jan 12, 2017 at 7:29
Sergiy Kolodyazhnyy
16.9k12 gold badges58 silver badges111 bronze badges
Here is an ugly sed way just for fun
sed '2,$ s/[^ ]\+/& &/; 2,$ s/ /\n/' file | paste -d ' ' - - abc 123 abc abc 789 bcd bcd 456 acb acb 135 Explanation
2,$from the second line to the lasts/[^ ]\+/& &/double the first set of non whitespace characters;separates commands, like in the shells/ /\n/replace the first space with a newlinepaste -d ' ' - -stick this mess together (append second line to third, fourth line to third, etc)
- 1Alternatively you could use
sedon its own withoutpaste:sed -r 'N;s/\n(\w+)/\1&/;P;D' somefile.txtDigital Trauma– Digital Trauma2017-01-13 00:19:48 +00:00Commented Jan 13, 2017 at 0:19 - 1If you write
sedprograms for fun, then perhaps you should have a go at code-golf ;-)Digital Trauma– Digital Trauma2017-01-13 00:34:20 +00:00Commented Jan 13, 2017 at 0:34 - 1@DigitalTrauma she's been on code-golf for 2 month already ;)Sergiy Kolodyazhnyy– Sergiy Kolodyazhnyy2017-01-13 03:43:34 +00:00Commented Jan 13, 2017 at 3:43
In my opinion the simplest and most readable approach is:
- extract the first column (
cut) - delete the first line from your extracted column (
tail) - paste this column into your source file (
paste)
Example: your sample inpult file:
abc 123 abc 789 bcd 456 acb 135 Then run the following command in a terminal
cut -d' ' -f1 in.txt | tail -n +2 | paste -d' ' file - Output:
abc 123 abc abc 789 bcd bcd 456 acb acb 135 The structure behind this solution differs from the given answers. No need of conditions, loops or regular expression.
