Shell Programming and Scripting

I would like to write a script that takes something like this:

Code:

Line 1
Line 2
Line 3
Line 4
[Blank line]
Line 6
Line 7
Line 8
Line 9
[Blank line]
etc....

And makes it to look like this:

Code:

Line 1|Line 2|Line 3|Line 4|[Blank line, if it's easier]
Line 6|Line 7|Line 8|Line 9|[Blank line, if it's easier]

etc.

I would think it would be fairly easy to use two loops, one for the the 5-line pattern and one for each set of 5 lines. However, I am having trouble planning out the code because if I create a loop within the loop, I have problems figuring out how to control which line to read. Any suggestions?
Perhaps have a loop that would create 1 super long line but insert a counter to insert a line break every 5 lines?

You could use awk to do this:

Code:

awk '{$1=$1}1' FS="\n" RS= OFS=\| infile

Or with paste:

Code:

paste -d\| - - - - - < infile

Can you explain what each component of the awk stmt means?
I am getting an error message:

Code:

awk: syntax error at source line 1
context is
     >>> {$1=$1}1 <<<
awk: bailing out at source line 1

(This was issued as a command, not as part of a script)

The paste command worked perfectly...actually too perfectly...because I found out that there are a few anomalies in the line sequences. Towards the end of the 50,000 lines in the file, the pattern changes to this:

Code:

Line A
Line B

Line C
Line D

And then it reverts back to the original pattern after those 2 sequences. I don't know why that's the case, but it's there.

You don't happen to know of any way to avoid that pothole, do you? (This is more wishful thinking)

Thanks for your help.

---------- Post updated at 12:31 PM ---------- Previous update was at 12:28 PM ----------

Oh I think if I can remove all blank lines in the file, then do a 4-line sequence, that should work as the 2-line sequences were actually halves of the 4-line sequence. So in light of that, how can I remove all blank lines from a file?

The awk commands are probably because you are on Solaris, if so try nawk instead of awk. The awk solution uses blank lines to align so your two anomalies will come out as:

Code:

Line A|Line B
Line C|Line D

But this should avoid getting out of sync for the rest of the file which is probably safer in case of other issues like this at a later date.

If you want to stick with paste you can use grep -v "^$" infile > outfile to remove blank lines, and don't forget to remove one of the dashes from the paste arguments to reduce to a 4-line sequence.

If not all empty lines are completely empty but some may contain spaces then instead of grep you could use:

Code:

awk NF file

You could then try running it through paste like Chubler suggests:

Code:

awk NF infile | paste -d\| - - - -

--
On Solaris < version 11 use /usr/xpg4/bin/awk or nawk rather than awk