sed adding an extra \n ???

kaz2100 · 04-10-2008, 11:10 AM

Hya,

Problem is it looks like that a sed script is adding an extra \n.

Background.
I have a bunch of files, format looks like

line 1: definiton
line 2-end: data (only [a-zA-Z]

and when data is bad, second line contains several X's only.

In short
good data files looks like

Code:

:This is an example of good file.
ADVRGBFGTHTFGRGRGETYHSBSHT
FVGRGBFBRHHGDFGHRHHFBDFGGD
...
SDFSDFSAFAS

bad data looks like

Code:

:This is an example for bad file.
XXXXX

and number of X is undefined (mostly under 20).

I wrote a sed script (filter) to eliminate bad data file, which looks like

Code:

#!/bin/sed -nf 
1 H
2 {     H
        s/^X*$/NG/
        T n
        Q
        :n 
        x
        p
        }
3,$ {   p
        }

When numerous input files are piped into the script above, it looks like that an additional \n between files is added.

I need to find where bug is...

I appreciate any help.

Happy Penguins!

osor · 04-10-2008, 03:06 PM

If you want to prevent a leading newline, you can just change:

Code:

1 H

to

Code:

1 h

ghostdog74 · 04-10-2008, 07:58 PM

you can use an alternative that keeps it simple.

Code:

awk 'NR==2 && /^XXXXX/{next}1' file