hi everybody
I'm new to shell scripting
i need to copy some lines in a file to another existing file.
my question is how to use commands of vi in the shell script ( copy command yy or yG, substitute command :s ....)


thanks in advance

Hi zoghbi,

You should use the cat, head, tail, awk commands for that purpose. VI is an interactive program, not suited for scripting. Check out the manual pages.

HTH

If you want to do this while you are working in vi/vim, why use a shell script when you could more easily use a vi/vim script instead? If you want it done by a shell script, then presumably you can programmatically identify/modify the lines of interest with tools like grep, awk, regular expressions etc, so why involve vi/vim?

dear porphyry5
thanks for your advice,
I need to use shell scripting because i have a lot of files (100 differenrs files) to effect the same function (copy the lines from the word 'system' (for example) to the end of each file(all the files contain this word), then past all this lines in another file)
i hope that you understand what i mean ( excuse me i 'm not good in English)

if you have an idea to do this work, please give us
thanks in advance

You can use "sed -n '/system/,$p' infile >> outfile" to append to another file. The range of lines printed are from one containing "system" to the end of the file.

Put it in a loop where "infile" is a variable, and you are set.

I'd highly recommend the book "Sed & Awk" by O'Reilly. Also, gawk has a manual, "Gawk: Effective Awk Programming", which is a pretty good reference manual for awk.

1 members found this post helpful.

thanks very much Mr jschiwal, your command work very well, thanks millions times.
i have another question:
if the word 'system' ( our example) is repeated twice in a file, and we would copy the lines from the second mention of this word to the end of the file, how should we do?

There are two ways offhand I can think of using sed.
This uses a subrange inside the main range. It will only work if you have exactly two lines with "system" in them.

A better way would be to append all lines to the buffer inside the /system/,$ range. But if such a line has another "system" in it, save the line, instead of appending, which will discard previously saved lines. On the last line, `$', append; recall the buffer; and print.

I'm not at my computer, and typing code is too painful on a small tablet, so I'll leave it as an exercise for now.

Another way would be to use ``tac'' to print the file out in reverse order. Pipe the output to sed.
tac file | sed '/system/{p;q}' | tac >> outfile

This will fail if the word ``system'' doesn't exist at all.