[RLUG] Re: quick question

[Grant Kelly]

I did play with the tail command, but as mentioned, you still have to
know the total number of lines in the file. `wc -l` took about 5
minutes to calculate this number, and then you've still got to run it
through tail.

Also, I wanted to remove the lines rather than ignore them so that, if
needed, I could reuse the file or pass it on to others without them
having to go through the same problem I did.
Furthermore, the CREATE TABLE commands were preceeded by something
like IF EXISTS DROP table.  I had already created the tables and
changed them slightly, so I didn't want them re-created.  sed was the
quickest and simplest tool for the job in this case.

So much for a "quick question".... over 30 messages in this thread!

Grant


On 6/30/06, James Washer <washer at trlp.com> wrote:
> Jeezus...
>
> YES... the tail command gives the last "xx" lines of the files.. but without counting the lines of the 2+GB file, one has no idea how many lines that is. Counting the lines of an arbitrarily large file, when there is no need to do so, is far from efficient.
>
> Further try "cat < head -3 /etc/passwd" and report back on the shell error you receive. That's just not legal shell syntax.
>
>
>
> On Fri, 30 Jun 2006 07:02:00 -0700 (PDT)
> Bill Cunningham <axoibc at yahoo.com> wrote:
>
> >
> >  Actually the tail command gives the last whatever:
> >
> > #!/usr/bin/perl -w
> >
> > $filename = "reallybighonknfile.sql";
> >
> > #number of lines in the entire file
> > $count = `wc -l $filename`;
> >
> > @output = `tail -l $count-300 $filename`;
> >
> > #@output has the part of the file you'll need to feed whatever you want.
> >
> > Oh and the "<" was to redirect the sql commands into the mysql process (with missing parameters for database and whatnot) which is what the original post asked for:
> >
> >
> > Bill Cunningham
> >