Shell Programming and Scripting

Hi Friends,

I looked up online, but couldn't figure out a proper solution.

I have an input file where the columns are separated by multiple spaces and the column content is separated by single space.

For example,

Code:

Chr1   hello world    unix is fun

In the above example, chr1 is first column, hello world is second column, unix is fun is third column. There are 3 spaces between first and second column, and 4 spaces between second and third columns. My required output is to replace single spaces in the column content only but not the separating space counts.

So, output will be

Code:

Chr1   hello_world    unix_is_fun

Please note that I have multiple columns.

Try:

Code:

sed 's/\([^ ]\) \([^ ]\)/\1_\2/g' file

for gnu flavor sed, try also:

Code:

sed 's/\b \b/_/g' file

GNU and some other sed understand the older left \< and right \> word boundaries

Code:

sed 's/\> \</_/g' file