UNIX for Dummies Questions & Answers

Hi:

it seems very strange.

there is a file with multiple lines. After I squeezed out the consecutive blank lines (and some other text processing), somehow the
sed '/\n/! d' file
can not generate any output, as if it can't find any line with newline.

the file is has many lines, so how could this be possible?

Thanks.

Don't understand your post.
Please post.
1) Sample input. In this case displayed in both characters and hexadecimal.
2) Expected output. Displayed in both characters and hexadecimal.


Btw: The "sed" command posted is just a syntax error on my systems.
Hard to guess what it was intended to do.

let's say the file "foo.txt" is:

hohoho
hahaha

the command is

sed '/\n/ d' foo.txt

The output is

hohoho
hahaha

the command is trying to delete all lines with the newline character. so all lines should be filtered out and no output should be generated at all.

But it seems that the sed did not find any match for '\n', which is strange.

There are should not any syntax errors.

Thanks.

you can try

Code:

sed ':a;N;$!ba;s/\n/ /g' file_name

Not strange. Newline is the record terminator in unix text files. sed works on text file format records and the record terminator is not loaded into the pattern space. In sed the end of the record is known as $.

Most of your previous posts have been about "csh" which is rarely used. Are you using csh in this post?

What are you actually trying to do?

Hi Methyl:

Thanks for the answer, very helpful to me to understand the sed better.

I guess I am confused somewhat by the fact that "sed 's%\n%%' fl" seems to work fine. So how come sed can match '\n' in the above "search and replace" structure, but not in a simpler '/\n/ d' structure?

my effort is more towards understanding sed than getting actual imminent work done.

I have to use csh not by my choice. Our EDA environment is setup that way. It actually is tcsh.

Thanks.

sed traditionally deals with data one line at a time and never matches newlines. It's odd that your version of sed ever matches a newline, mine certainly doesn't! If you want to delete newline characters, tr -d '\n' would work better: it deals with bytes, not lines.