How to use vi to edit only the top 10 lines file

[alexsqlforums]

I have a 6.4 GB .sql file and i need to modify only the top 10 line portion of that file.
How can i only open the top 10 lines of this file without fully opening it?
OR
How can i search and replace a certain string like "Create Database mydb1" and comment it out without fully opening it?

Thank you
Any help appreciated.

[smolloy]

You could use a combination of awk and sed. Use awk to print out the top 10 lines to another file for editing (you could even use awk to edit them in transit to the new file). Then use sed to delete those lines in the original file, and then, finally, concatenate the edited lines onto the top of the old file.

[danielsmw]

I'm not an expert with sed, but you may want to look into it a bit. This sounds like something you could use it for. Another possibility is to actually split it into two files:

head -n 10 file.sql > fileTop
wc -l file.sql
tail -n (number of lines - 10) > fileBottom

That wc line is just there to figure out the number of lines in the file. Then you could edit fileTop and when you're done, just put the files back together. This is by no means an elegant solution, and I'm sure there are better answers, but that's just one way of approaching it.

Another way could be to write a quick Perl script. Just open the the file, use your substitution for the first few lines ($line =~ s/replacingthis/withthis/ if $linenumber < 10, and then just print the rest of the lines as well.

But hopefully, someone has a better solution, because I want to see it too!

[danielsmw]

oops, and smolloy posts a better solution as I type. =)

[smolloy]

oops, and smolloy posts a better solution as I type. =)

I'm not sure my solution is better. Your idea of using "head" to extract the top 10 lines is more elegant than awk (although you'd have to pipe it to sed if you wanted to edit the lines in transit as I suggested).

Maybe a combination of our ideas is the best

[alexsqlforums]

Thanks guys for pointing out to awk and sed.
After a little research i was able to replace the string by using this command.

sed 's/CREATE DATABASE/#CREATE DATABASE/' abc.sql | tee abc1.sql

[smolloy]

I was going to suggest that, but then I noticed that it was a very large file. Didn't it take a long time to pass that monster through sed?

[alexsqlforums]

it did
Now i'm trying to implement awk and cut/edit/paste back to top of file.

I'll figure it out soon

thank you