brooks@sundance7.dab.ge.com (Stephen Brooks) (05/04/91)
I have a sed question for you gurus: assume I have a file of the form
name1; A1.1 B10.20 C100.300,
D101.25 E202.50,
F300.7
name2; Z44.33 Y409.45
name3; X777.77 W6.6,
V32.15
where a comma (,) represents a continuation character. I need to "massage"
this into something which looks like this:
name1; A1.1 B10.20 C100.300 D101.25 E202.50 F300.7
name2; Z44.33 Y409.45
name3; X777.77 W6.6 V32.15
Can I do this in sed? I've tried using the N command in sed, but the problem
with it is I don't see how to handle multiple continuation lines without using
multiple N commands. Is this possible? What am I missing?
I currently have to do something like this (in sed):
/,$/ N find "," at end of line, append next line to input pattern
s/\n// delete embedded <CR>
/,$/ N do it again (and again and again...)
s/\n//
.
.
.
Thanks!
--
%% Stephen (Steve) M. Brooks %% brooks@ge-dab.ge.com %%
%% GE Simulation & Control Systems %% ...!uunet!ge-dab.ge.com!brooks %%
%% P.O. Box 2825, Rm. 1370 %% ...!uunet!sunny.dab.ge.com!brooks %%
%% Daytona Beach, FL 32115-2825 %% voice: (904) 239-4855 %%rouben@math16.math.umbc.edu (Rouben Rostamian) (05/04/91)
In article <1991May3.213550.17246@ge-dab.GE.COM> brooks@sundance7.dab.ge.com (Stephen Brooks) writes: > I have a sed question for you gurus: assume I have a file of the form > >name1; A1.1 B10.20 C100.300, > D101.25 E202.50, > F300.7 >name2; Z44.33 Y409.45 >name3; X777.77 W6.6, > V32.15 > >where a comma (,) represents a continuation character. I need to "massage" >this into something which looks like this: > >name1; A1.1 B10.20 C100.300 D101.25 E202.50 F300.7 >name2; Z44.33 Y409.45 >name3; X777.77 W6.6 V32.15 > >Can I do this in sed? Yes. Here it is: sed -n ' :loop /,$/{N bloop } s/,\n//g s/ */ /g p' <inputfile -- Rouben Rostamian Telephone: (301) 455-2458 Department of Mathematics and Statistics e-mail: University of Maryland Baltimore County bitnet: rostamian@umbc.bitnet Baltimore, MD 21228, U.S.A. internet: rouben@math9.math.umbc.edu
merlyn@iwarp.intel.com (Randal L. Schwartz) (05/05/91)
In article <1991May3.225202.29639@umbc3.umbc.edu>, rouben@math16 (Rouben Rostamian) writes: | In article <1991May3.213550.17246@ge-dab.GE.COM> brooks@sundance7.dab.ge.com (Stephen Brooks) writes: | > I have a sed question for you gurus: assume I have a file of the form | > | >name1; A1.1 B10.20 C100.300, | > D101.25 E202.50, | > F300.7 | >name2; Z44.33 Y409.45 | >name3; X777.77 W6.6, | > V32.15 | > | >where a comma (,) represents a continuation character. I need to "massage" | >this into something which looks like this: | > | >name1; A1.1 B10.20 C100.300 D101.25 E202.50 F300.7 | >name2; Z44.33 Y409.45 | >name3; X777.77 W6.6 V32.15 | > | >Can I do this in sed? | | Yes. Here it is: | | sed -n ' | :loop | /,$/{N | bloop | } | s/,\n//g | s/ */ /g | p' <inputfile And in Perl... (see, I waited for the sed solution to be posted this time :-): perl -pe 'while (/,\s*$/) {$_ .= <STDIN>; s/,\s*\n\s*/ /;}' <in >out (which means, while you see a comma at the end of the line (ignoring trailing whitespace), append the next input line, and replace the comma, optional whitespace, newline, and more optional whitespace with a single space.) print "Just another Perl hacker," -- /=Randal L. Schwartz, Stonehenge Consulting Services (503)777-0095 ==========\ | on contract to Intel's iWarp project, Beaverton, Oregon, USA, Sol III | | merlyn@iwarp.intel.com ...!any-MX-mailer-like-uunet!iwarp.intel.com!merlyn | \=Cute Quote: "Intel: putting the 'backward' in 'backward compatible'..."====/
usenet@carssdf.UUCP (John Watson) (05/06/91)
> perl -pe 'while (/,\s*$/) {$_ .= <STDIN>; s/,\s*\n\s*/ /;}' <in >out > /=Randal L. Schwartz, Stonehenge Consulting Services (503)777-0095 ==========\ In this example, don't you have to set $* =1; to get the s/,\s*\n\s*/ to match. John Watson
merlyn@iwarp.intel.com (Randal L. Schwartz) (05/06/91)
[I tried mailing. It bounced. Where the heck is carssdf? And yes, further discussion to comp.lang.perl, please...] In article <304@carssdf.UUCP>, usenet@carssdf (John Watson) writes: | > perl -pe 'while (/,\s*$/) {$_ .= <STDIN>; s/,\s*\n\s*/ /;}' <in >out | | In this example, don't you have to set $* =1; to get the s/,\s*\n\s*/ to | match. | | John Watson No. $* is only for selecting whether ^ and $ match a begin/end-of-line within the buffer. \n will always match itself, as will \s match newline (and the other whitespace). The \n is in there to ensure that I kill the *right* comma and whitespace (which the sed solution doesn't guarantee). Just another Perl hacker (but of course, you knew that), -- /=Randal L. Schwartz, Stonehenge Consulting Services (503)777-0095 ==========\ | on contract to Intel's iWarp project, Beaverton, Oregon, USA, Sol III | | merlyn@iwarp.intel.com ...!any-MX-mailer-like-uunet!iwarp.intel.com!merlyn | \=Cute Quote: "Intel: putting the 'backward' in 'backward compatible'..."====/