Problem
I want to convert all of the tabs in my text file to spaces. Reasons for this include improper indentation when printing, a discrepancy between the operating environment and the file tab size, files included in e-mail, and so on.
Cause
If you insert text in vi(C) while the 'ai' (autoindent) is set, then embedded tab characters may be put in the file. It is also possible that these tab characters were explicitly entered by you or some application program.
NOTE: Many UNIX/XENIX utilities that process files are
line-oriented, meaning that the lines are <NL>-terminated.
Also, most assume that the the operating environment default
is for tabs to be equivalent to 8 characters. If the file you
are converting had a different assumption, then the conversion
program will need an additional option. When applicable, the
examples in this document are given for the default tab size
of 8, and for a tab size of 4. Solution
You can convert embedded tabs to spaces either I) at the shell
command line or II) from within the vi editor. There are
separate recommended solutions for A) SCO OpenServer Release 5.0,
B) SCO UNIX (prior to release 5.0) and C) SCO XENIX. Additionally,
the appendix contains a sample shell script front-end for converting
tabs to spaces.I. Command-line utilities:
--------------------------
A. SCO OpenServer Release 5.0The expand(C) filter can be used to expand tabs to spaces as follows:
tabsize=8
---------
cat filename.old | expand > filename.newor the more effecient...<filename.old expand >filename.newtabsize=4
---------
<filename.old expand -t4 >filename.new(where filename.old is the name of the file containing tabs and filename.new is the name of the converted file)
NOTE: The expand utility is not included in SCO Operating Systems prior to OpenServer Release 5.0.
B. SCO UNIX, Open Desktop, Open Server (before release 5.0)
The col(C) filter can be used to expand tabs to spaces as follows:
tabsize=8
---------
cat filename.old | col -x > filename.newor the more effecient...<filename.old col -x >filename.newtabsize=4
---------
pr -te4 filename.old > filename.new(where filename.old is the name of the file containing tabs and filename.new is the name of the converted file)
NOTE: The col utility is not included with SCO XENIX. It is included with the SCO XENIX Text Processing package.
C. SCO XENIXThe pr(C) filter can be used to expand tabs to spaces as follows:
tabsize=8
---------
pr -te filename.old > filename.newtabsize=4
---------
pr -te4 filename.old > filename.new(where filename.old is the name of the file containing tabs and filename.new is the name of the converted file)
NOTE: The pr utility is included with SCO UNIX, but it is
better to use col with UNIX. The UNIX pr utility
expands the first tab found on the first line of each page
by one too many spaces. This problem can be alleviated by
using the flag "-l10000" ("10000 lines per page").II. Within 'vi' editor:
-----------------------
You can also convert tabs to spaces when editing the text
file from within the vi editor. First, you may want to
see whether there are any tabs embedded in the file.
Enter the following sequence:<ESC> 1 G /<TAB> <ENTER>If you get a message such as "Pattern not found" or "Address search hit BOTTOM without matching pattern", that means that there are no <TAB> characters embedded within the file.
Otherwise, vi did find tab character(s) in the file. So
next, to actually convert the tabs to spaces, bring up the
'ex' command line by typing colon (:).Then at the ex command line ':' prompt, type:
A. SCO OpenServer Release 5.0tabsize=8 tabsize=4
--------- ------------
%!expand %!expand -t4B. SCO UNIX, Open Desktop, Open Servertabsize=8
---------
%!col -xC. SCO XENIXtabsize=8 tabsize=4
--------- ---------
%!pr -te %!pr -te4NOTE: If after the above command the file appears corrupted,
simply press "u" to undo the last operation. See also
man pages: col(C), pr(C), vi(C), umask(C), login(M),
umask(S), expand(C)Appendix
========
This shell script will convert a file with tabs into one
with its tabs replaced by spaces. Save this to a file
named "exp8" under your local bin directory, change mode
to 0755, and then link the file to "exp4".-- cut here --- cut here --- cut here --- cut here --- cut here--
:
# 'exp4/exp8: expand tab4/8 to spaces'
# input: filename with embedded tabs
# output: filename with tabs converted
# syntax: exp8 <filename>
# exp4 <filename>
#trap "echo 'no change'; exit 0" 1 2 3 15 trap "rm -f /tmp/exp$$" 0
if [ $# -ne 1 -o ! -w "$1" ]; then
echo "usage: $0 filename"
exit 1
fiFULLNAME=$1
if [ "`basename $0`" = exp4 ]; then
SPACE=4
else
SPACE=8
fipr -t -e$SPACE $FULLNAME > /tmp/exp$$ &&
cp /tmp/exp$$ $FULLNAME-- cut here --- cut here --- cut here --- cut here --- cut here--
NOTE: This command will also process a filename with spaces within the
filename:eg, here's a sample file_list:data/cyrus/imap/user/falkirk_hiredrive/Sent Items/core data/cyrus/imap/user/fiat_fleet/Sent Items/core data/cyrus/imap/user/fpaton/Sent Items/core data/cyrus/imap/user/gblack/Memo Scott Willis/core data/cyrus/imap/user/yrae/Sent Items/core data/cyrus/imap/user/gizat/Sent Items/core data/cyrus/imap/user/gjohnston/Peugeot, PFS &- Branch/core data/cyrus/imap/user/glasgow_daewoo/AC Memos/core data/cyrus/imap/user/glasgow_hyundai/AC Memos/core data/cyrus/imap/user/glasgow_hyundai/Sent Items/core data/cyrus/imap/user/glasgow_hyundai/Martin Geary/core data/cyrus/imap/user/glasgow_hyundai/Sandy Biggart/core data/cyrus/imap/user/jbertram/from franchise/core data/cyrus/imap/user/grangemouth_renault/Sent Items/core data/cyrus/imap/user/citroendev/Sent Items/core data/cyrus/imap/user/gtraining/Sales Academy/core data/cyrus/imap/user/gtraining/Management Academy/core data/cyrus/imap/user/hhamilton/Junk E-mail/core data/cyrus/imap/user/inverness_citroen/Sent Items/core data/cyrus/imap/user/janderson/Sent Items/core data/cyrus/imap/user/prebooking/Sent Items/core data/cyrus/imap/user/jclarke/Sent Items/core data/cyrus/imap/user/lhenderson/Sent Items/core
and to remove them:# ( IFS=$'\n' ; rm $(cat file_list ) )