Shell Programming and Scripting

I have a data file where three data sets are written in three columns. Can I increase the space between the columns without reading them?


Also can I insert particular patterns, say comma between 1st and 2nd column and colon between 2nd and 3rd column?

This is some simple examples:

Code:

$ cat f1
123 abc 456
xyz 789 efg

# Replace space between 1st and 2nd column with 3 spaces
$ sed 's# #   #' f1
123   abc 456
xyz   789 efg

# Replace space between all columns with 3 spaces
$ sed 's# #   #g' f1
123   abc   456
xyz   789   efg

# Replace space between 1st and 2nd column with a comma
$ sed 's# #,#' f1
123,abc 456
xyz,789 efg

# Replace space between all columns with a comma
$ sed 's# #,#g' f1
123,abc,456
xyz,789,efg

Not unless the data from all three data sets is a constant number of bytes in each line in the file AND you are replacing a constant number of bytes with the same number of bytes that increases the space between them (such as by replacing a space character with a tab character). Under these circumstances you could write a C or C++ program to just write replacement characters into fixed positions in the file. (Even then the OS would have to read the blocks containing the bytes that surround the bytes you're replacing; but your program wouldn't have to directly read the file.)

If you want to insert a comma or a colon AND increase the space between fields, you cannot do so without reading the data unless you'd be replacing a multi-byte character of length 2 or more with a single-byte comma or colon and another single-byte character (such as space or tab) and enough other single-byte characters to exactly fill the bytes occupied by the multi-byte character you'd be replacing.

If you're willing to read the file, there are thousands of ways to change field separators between fields if you'll give us an accurate description of your input file(s) including what the input field separators are and describe what field separators you want to replace them in your output.

I get only change in space between 1st and 2nd column. Could you explain what is happening?

I also believe that there must be a smart 'awk'-way to do these.

Two awk examples:

Code:

# Separate columns with 3 spaces
$ awk -v OFS='   ' '{$1=$1};1' f1
123   abc   456
xyz   789   efg

# Separate columns with a comma
$ awk -v OFS=',' '{$1=$1};1' f1
123,abc,456
xyz,789,efg

What about my original second question:
comma between 1st and 2nd column, and colon between 2nd and 3rd column ?

Code:

$ awk '{ print $1","$2":"$3 }' f1
123,abc:456
xyz,789:efg