Shell Programming and Scripting

Hi.

How do I convert the Windows new line into a known format to Unix?

Please see the attached file.

In UNIX it appears only one single line.

Please help.

Thank you.

Try:

Code:

tr '\r' '\n' < INPUT_GIANT_JAN2016.txt > FIXED_INPUT_GIANT_JAN2016.txt

Thanks a lot!

It works!

But what about applying the same command that won't affect the file that is already OK, like this one..

I've tried the command but it will create a blank lines between the record.

Is it possible?

In other words, I wouldn't know if the files are OK or not, so need to apply general command to all files.

Thank you.

UNIX format text files have a newline character as a line terminator. DOS format text file have a carriage-return character immediately followed by a newline character with the combination of those two characters being the line terminator.

The first file you asked about (INPUT_GIANT_JAN2016.txt) was not a normal DOS format text file. It only had the carriage-return character as the line terminator (except for the last line in the file that had both carriage-return and newline characters). So to fix it, all we had to do was change the carriage-return characters to newline characters. (That left you with a single empty line at the end of the file after tr changed all of the carriage-return characters to newline characters.)

The second file (INPUT_GUARDIAN_JAN2016.txt) is a normal DOS format text file. So instead of changing carriage-return characters to newline characters, we just need to remove the carriage-return characters:

Code:

tr -d '\r' < INPUT_GUARDIAN_JAN2016.txt > NEW_INPUT_GUARDIAN_JAN2016.txt

If you only have files in these two exact formats, you could handle both with:

Code:

tr '\r' '\n' < input_file | grep -v '^$' > fixed_file

This won't work for general text files (which frequently have empty lines separating paragraphs), but it should work for the samples you have shown us so far.

Thanks a lot Don!

It works.

But I just wondering, both type of files (Unix line and Windows line), the final output would have less 1 byte.

But the content look similar.

Is that expected?

Thank you.

Compare the two files with e.g. od or hexdump.