[bash] Shell scripting VI commands

Hi there,

I have the following problem. I need to copy across a tab-delimited text file from one server to another, strip the first two lines (which are a header of information) and then import into mysql.

I want this all this done in a shell script. I have solved the first issue (using ssh-keygen) and can copy to and from different servers.

However, I want (in my shell script) to open vi and delete two lines (so, `dd` twice). Can anyone help me with scripting vi commands? Every time I use Google with the search words "vi shell scripting" I get literally thousands of results of the like: "when I shell script I use vi" etc etc - no use at all.

Any help appreciated - even links to helpful websites . My shell is bash and my OS is Linux Red Hat.

Cheers,

- Tatlar

You can do it without using vi. I'm using wc, cut, dc and tail to do this. Promise me that you won't laugh at this approach .


I'm assuming that the filename to process is passed as an argument to the script. e.g.
./clipit.sh file_to_process.txt

HTH

Thanks for the help - I won't laugh if it solves my problem :P

- T

hey Scorpions4ever,

thanks buddy! - that worked like a charm and has made my life infinitely better

i owe you one.

- tatlar

tatlar...

Another way to get the lines count would be just
cat filename | wc -l

I'm not sure why Scorpions4ever used cut, since wc -l gives you a one-field answer, but if you needed to get *just* the arithmetic value, you could use this:
lines=$((`cat filename | wc -l`))

The rest of his post is excellent; I wasn't laughing at all.
;)

By the way, I've wondered a few times how to get vi commands executed in a shell script. if you find out, please let me know.

Scorpions4ever, your .sig is sweeeeet. That's awesome!

>> I'm not sure why Scorpions4ever used cut, since wc -l gives you a one-field answer

Executing wc -l filename on my RedHat Linux system outputs both the line count and the filename (i.e. two fields). That's why I used cut to chop out the line count field from the returned line.

HTH

heh...

My first response was

"WHAT?!??! That's so weird, because I've never heard of that before!"

But then I paid closer attention to your syntax. I hadn't realized that I've always used wc in the secondary position, never as primary:

cat .profile | wc -l

I just tested it by executing this:
wc -l .profile

...and it was just like you said:
48 .profile

...on Open BSD
...and Slackware
...and AIX
...and Solaris
...and Red Hat

so, it looks like I've proven once again (I've done it a lot) that you can post a comment THINKING you're making a correction, but really you're just ignorant. I was talking just today with someone about learning constantly.

Hmmmm. Wonder who that could've been....
;)

-zedmelon

to make things painfully apparent, since at the moment, they're painfully *not* apparent, I, zedmelon, was the ignorant party I referenced in the last paragraph saying something like, "...you can be ignorant, yet think you're making a good point..."

Heh... no point in starting a flame war if I'm aiming the first flaming arrow at myself but simply miss.
;)

-zedmelon

One simple way that I have used vim in shell scripts was as follows (ok, I could have used sed/awk/etc, but I really like vi)
(ok, I don´t know how to use the others so I've used vi)

I was trying to change newlines to spaces and what I have done:

vim '+%s/\n/ /g' '+wq' file2edit.txt

I hope that could be useful

Regards,
Ailton Oliveira