Remove line feeds

Shell Programming and Scripting
1

Hi,

I have a fixed width flat file which has 1 as the first char and E as the last character. Some of the records have a carriage return /line feeds .
how do I remove them?

Let me know.

Thanks
VSK

2

You could use sed or the tr command to strip off the characters you don't want. Here is a simple tr example.

This would remove carriage returns:

% cat yourfile.txt | tr -d "\r"

This would remove line feeds:

% cat yourfile.txt | tr -d "\n"

This would remove both...

% cat yourfile.txt | tr -d "\n" | tr -d "\r"
3

Just out of curiosity, how would you do that using sed?

Thanks,
djp

4

But that will remove the \n and \r whcih are found in the end of each line and I do not want to do that

Following is the example:

1abcdef E
1ghijk E
1kmijol E
1kfcldsa
cdsc;mE
1kcdmlms E

Here want to remove the carrriage return from the last but one line, since we need 1 as the first char and E as the last char

5

I guess the newlines would be tricky :rolleyes: but the returns should be removable with..

cat infile.txt | sed 's/\r//g' > outfile.txt

Now you have me wondering.... :confused:

6

This might help

sed '/^$/d' infile.txt > outfile.txt

check it out.... :cool:

7

I think we misunderstood what the threadstarter wanted: The task is to concatenate all lines starting NOT with a "1" to the last line so that all lines start with a "1" and end with an "e":

source.file:
1 this line is ok E
1 This line starts ok
but continues E
1 here is the next line E

target.file:
1 this line is ok E
1 This line starts ok but continues E
1 here is the next line E

This can easily be accomplished by sed's "N" and "P" subcommands, which reads in the next line and print the whole pattern space respectively.

I could say "man sed", but I'd rather like to suggest reading Dale Doughertys fantastic book "sed & awk" from O'Reilly Publishing. He discusses exactly such cases as examples for setting up what he calls a "while..do loop in a sed-script".

bakunin

8

awk ' {
printf $0
while (length ($0) && substr ($0, length ($0), 1) != "E") {
getline; printf $0
}
if (length ($0) == 0) print "E"; else print ""
}' infile > outfile

9

gnu sed

sed -e ':a' -e 's/\r//g; /[^E]$/{N; s/\n//; ta}'