I am new to Linux and have a challenging task.
I have 3 data files, and need to do the following:
Go to line 31 of file 1, delete it
Read 1 line from file 2 and add in place of deleted line
Go to line 97 of file 1 delete it and then read the line 1 from file 2 and add in place of that deleted line in file 1.
The thing is also important to keep the same file i.e file , it is not to be changed.
I tried different versions of sed and perl, with buffer copying tricks but was not successful.
I am open for all suggestions and request the experts to give me suggestions.
I cannot find a reference to the 3rd file in your question, but if you mean replace line number 31 of file 1 with the 1st line of file 2, and replace line number 97 of file 1 with the 2nd line of file 2:
sed -i -e '30R f2
31d;96R f2
97d' f1
The new lines are important after f2 so sed knows that it is the end of the file name.
Note that the R command is a GNU extension, it is not standard.
Related
I just need to move a line up in sed. I can select the line with
sed -i '7s///'
I need to move line 7 up 2 lines so it will be line 5.
I can't find anything on the internet to do this without complicated scripts, I can't find a simple solution of moving a specific line a specific number of times.
seq 10|sed '5{N;h;d};7G'
when up to line 5 append next line(line 6) into pattern space then save them into hold space and delete them from pattern space; up to line 7 then append the hold space content("5\n6") behind the line 7; now, pattern space is "7\n5\n6";finally,sed will print the pattern space at the end of current cycle by default(if no "-n" parameter)
ed is better at this, since it has a "move" command that does exactly what you want. To move line 7 to be the line after line 4, just do 7m4. ed doesn't write the data back by default, so you need to explicitly issue a w command to write the data:
printf '7m4\nw\n' | ed input
Although it is perhaps better to use a more modern tool:
ex -s -c 7m4 -c w -c q input
I tried p4 diff2 command. Please help me in understanding the command's output. Specifically I want to understand the lines like following
1c1
3a4,5
6a9
13a17,18
2,7d1
It's pretty typical; see https://unix.stackexchange.com/questions/81998/understanding-of-diff-output for a very similar description.
Lines which were (a)dded, (c)hanged, or (d)eleted are prefixed by these entries.
The numbers are line numbers from either the source file or the target file.
So 1c1 means that line 1 in the source was changed, resulting in line 1 in the target.
And 13a17,18 means that lines 17-18 in the target were added after line 13 in the source.
And 2,7d1 means that lines 2 through 7 in the source were deleted before line 1 in the target.
I just need to move a line up in sed. I can select the line with
sed -i '7s///'
I need to move line 7 up 2 lines so it will be line 5.
I can't find anything on the internet to do this without complicated scripts, I can't find a simple solution of moving a specific line a specific number of times.
seq 10|sed '5{N;h;d};7G'
when up to line 5 append next line(line 6) into pattern space then save them into hold space and delete them from pattern space; up to line 7 then append the hold space content("5\n6") behind the line 7; now, pattern space is "7\n5\n6";finally,sed will print the pattern space at the end of current cycle by default(if no "-n" parameter)
ed is better at this, since it has a "move" command that does exactly what you want. To move line 7 to be the line after line 4, just do 7m4. ed doesn't write the data back by default, so you need to explicitly issue a w command to write the data:
printf '7m4\nw\n' | ed input
Although it is perhaps better to use a more modern tool:
ex -s -c 7m4 -c w -c q input
I copy pasted some enum values from my IntelliJ IDE in windows to notepad, saved the file in a shared drive, then opened it up in a linux box. When I did cat -A on the file it showed something like:
A,B,C,^M$
D,E,F,^M$
G,H,I,^M$
After searching around I figured that ^M is the carriage return and $ means the last line of the file. I'm just puzzled at how this file is able to have multiple $'s.
From man cat on my GNU box:
-A, --show-all
equivalent to -vET
(snip)
-E, --show-ends
display $ at end of each line
Thus, there are multiple $s because there are multiple lines, each with an end.
$ is the end of line marker with cat -A, not end of file.
This is indicating the file has Windows-style line endings (carriage return followed by line feed) and not Unix-style (only line feed).
(You can convert text files from one format to the other using the programs dos2unix or unix2dos.)
I'm having trouble applying a patch to my source tree, and it's not the usual -p stripping problem. patch is able to find the file to patch.
Specifically, my question is how to read / interpret the .rej files patch creates when it fails on a few hunks. Most discussions of patch/diff I have seen don't include this.
A simple example:
$ echo -e "line 1\nline 2\nline 3" > a
$ sed -e 's/2/b/' <a >b
$ sed -e 's/2/c/' <a >c
$ diff a b > ab.diff
$ patch c < ab.diff
$ cat c.rej
***************
*** 2
- line 2
--- 2 -----
+ line b
As you can see: The old file contains line 2 and the new file should contain line b. However, it actually contains line c (that's not visible in the reject file).
In fact, the easiest way to resolve such problems is to take the diff fragment from the .diff/.patch file, insert it at the appropriate place in the file to be patched and then compare the code by hand to figure out, what lines actually cause the conflict.
Or - alternatively: Get the original file (unmodified), patch it and run a three way merge on the file.
Wiggle is a great tool for applying .rej files when patch does not succeed.
I'm not an expert on dealing with patch files, but I'd like to add some clarity on how to read them based on my understanding of the information they contain.
Your .rej files will tell you:
the difference between the original and the .rej file;
where the problem code starts in the original file, how many lines it goes on
for in that file;
and where the code starts in the new file, and how many lines it goes on for in that file.
So given this message, noted in the beginning of my .rej file:
diff a/www/js/app.js b/www/js/app.js (rejected hunks)
## -4,12 +4,24 ##
I see that for my problem file (www/js/app), the difference between the original (noted as a/www/js/app.js on the first line) and the .rej file (noted as b/www/js/) starts on line 4 of the original and goes on for 12 lines (the part before the comma in ## -4,12, +4,24 ## on line two), and starts on line 4 of the new version of the file and goes on for 24 lines (the part after the comma in ## -4,12, +4,24 ##.
For further information, see the excellent overview of patch files (containing the information I note above, as well as details on lines added and/or between file versions) at http://blog.humphd.org/vocamus-906/.
Any corrections or clarifications welcome of course.