Get first N lines of output from a pipe operation

Question

I'm using a piped command to migrate a big production DB from one host to another using this command:

mysqldump <someparams> | pv | mysql <someparams>

And I need to extract the line 23 (or let's say the first X lines) (saved as file or simply in bash output) from the SQL passing from one server to another.

What I've tried:

• Concatenate in output less, at least to see the output scrolling, but no luck

  mysqldump <someparams> | pv | mysql <someparams> | less

• Read about sed, but it's not useful to me

• Using head to write to a file, but it is empty

mysqldump <someparams> | pv | mysql <someparams> | head -n 25 > somefile.txt

The only requirement I have is that I cannot save this .sql file.

Any idea?

Thanks

Answer 1

With zsh

mysqldump <someparams> |
  pv > >(sed '22,24!d' > saved-lines-22-to-24.txt) |
  mysql <someparams>

With bash (or zsh):

mysqldump <someparams> |
  pv |
  tee >(sed '22,24!d' > saved-lines-22-to-24.txt) |
  mysql <someparams>

(though beware that as bash doesn't wait for that sed process, it's not guaranteed that saved-lines-22-to-24.txt will be complete by the time you run the next command in the script).

Or you could have sed to the writing:

mysqldump <someparams> |
  pv |
  sed '22,24 w saved-lines-22-to-24.txt' |
  mysql <someparams>

To have it as output, with zsh:

{mysqldump <someparams> |
  pv > >(sed '22,24!d' >&3) |
  mysql <someparams>} 3>&1

or bash/zsh:

{ mysqldump <someparams> |
  pv |
  tee >(sed '22,24!d' >&3)
  mysql <someparams>
} 3>&1