patkar@amide.ecn.purdue.edu (The Silent Dodo) (08/13/90)
I need to do the following in vi. I have a file containing multiple occurances of a single letter. A sample line is ccccoooossss I need to replace all quadruple occurance by a single letter. Thus the above line should become cos I tried something like :1,$s/\(a-zA-Z)\)\1\1\1/\1/g Vi does not allow \1 in the pattern to be searched. Any pointers will be appreciated. -- Anant (patkar@cn.ecn.purdue.edu)
meyering@cs.utexas.edu (Jim Meyering) (08/13/90)
In article <1990Aug12.194738.7902@ecn.purdue.edu> patkar@amide.ecn.purdue.edu (The Silent Dodo) writes: > I need to do the following in vi. [ Change text like "ccccoooossss" to "cos" ] > :1,$s/\(a-zA-Z)\)\1\1\1/\1/g [ doesn't work ] If your "in vi" requirement is strict, I'm not sure I can help. Otherwise, `sed' can do it. From in vi, :%!sed 's/\([a-zA-Z]\)\1\1\1/\1/g' works for me. (It looks like maybe you had an extra right paren and forgot the square brackets) Regards, -- Jim Meyering meyering@cs.utexas.edu
was@hp-lsd.COS.HP.COM (Bill Stubblebine) (08/14/90)
patkar@amide.ecn.purdue.edu (The Silent Dodo) > I need to do the following in vi. > ccccoooossss -> cos :%s/cc*/c/g :%s/oo*/o/g :%s/ss*/s/g
grahj@gagme.chi.il.us (jim graham) (08/14/90)
> I need to do the following in vi. >[ Change text like "ccccoooossss" to "cos" ] If you are trying to get rid of the attempt at BOLD text that nroff uses (i.e., backspacing and re-typing the same character over and over again), there is a better solution.... There is a program called ``pep'' that resides on a number of the machines that archive sources for UN*X, that will filter out all of this nonsense for you. If the original file is called foo.txt, you would do something like this: pep < foo.txt > foo.new mv foo.new foo.txt (if you like foo.new....) NOTE: I include the part about ``if you like foo.new'' because I'm not sure if you need to use just ``pep'' or ``pep -b''. I know one of those works perfectly (IF THIS IS WHAT YOU HAD IN MIND). Hope this helps..... Jim Graham
patkar@ecn.purdue.edu (Anant Y Patkar) (08/15/90)
In article <401@taumet.com>, steve@taumet.com (Stephen Clamage) writes: |> |:%s/cc*/c/g |> |:%s/oo*/o/g |> |:%s/ss*/s/g |> |> This then also changes |> 'accent' to 'acent' |> 'loop' to 'lop' |> 'ass' to 'as'. |> I wonder if that is what the original poster had in mind. Just to clear the confusion. What I needed was to change quadruple occurances of any letter to a single occurance. That means if I use individual substitute commands for each letter, I am stuck with 52 commands (some files I have need more commands to take care of some other characters). Also it should change only quadruple occurances. I don't want any spelling mistakes in the resulting file. The solutions I received my mail indicate that (possibly) there is no way to do this using the "ex" commands. You have to invoke some UNIX commands like 'sed' from inside vi to do this. Hope this helps! -- Anant (patkar@cn.ecn.purdue.edu)
dolf@idca.tds.PHILIPS.nl (Dolf Grunbauer) (08/16/90)
In article <14420003@hp-lsd.COS.HP.COM> was@hp-lsd.COS.HP.COM (Bill Stubblebine) writes: >patkar@amide.ecn.purdue.edu (The Silent Dodo) > >> I need to do the following in vi. >> ccccoooossss -> cos > >:%s/cc*/c/g >:%s/oo*/o/g >:%s/ss*/s/g Why not: :%s/cc*oo*ss*/cos/g ? I am missing something ?
steve@taumet.com (Stephen Clamage) (08/16/90)
was@hp-lsd.COS.HP.COM (Bill Stubblebine) writes: |> I need to do the following in vi. |> ccccoooossss -> cos |:%s/cc*/c/g |:%s/oo*/o/g |:%s/ss*/s/g This then also changes 'accent' to 'acent' 'loop' to 'lop' 'ass' to 'as'. I wonder if that is what the original poster had in mind. -- Steve Clamage, TauMetric Corp, steve@taumet.com
egt@hprnd.HP.COM (Eric Tausheck) (08/17/90)
> > I need to do the following in vi. > > ccccoooossss -> cos > > :%s/cc*/c/g > :%s/oo*/o/g > :%s/ss*/s/g > ---------- Wouldn't one of the following be more accurate? :%s/cc*oo*ss*/cos/g :%s/ccccoooossss/cos/g I had the need to do this kind of editing too until I learned about col -b. If what you started with had a bunch of back-spaces between the letters too (as in an nroff or man output), the filter "col -b" does a nice job of getting rid of all this junk for you.
grahj@gagme.chi.il.us (jim graham) (08/17/90)
In article <14420003@hp-lsd.COS.HP.COM> was@hp-lsd.COS.HP.COM (Bill Stubblebine) writes: >patkar@amide.ecn.purdue.edu (The Silent Dodo) > >> I need to do the following in vi. >> ccccoooossss -> cos > >:%s/cc*/c/g >:%s/oo*/o/g >:%s/ss*/s/g First off, if the sources I sent patkar@..... work, this is a dead issue. the intent was to get rid of the overstriking done by nroff (etc) for bold type. Second, and here's where I may simply have more to learn about vi (which is the case all the time.....), with your search/replace commands above, what happens with strings such as "success", "pool", and so on? I guess the real question should be, does the "*" tell vi that you're talking about strings of more than 2 of that character? Thanks, Jim Graham PS: RTFM isn't a valid answer, since I didn't find the answer there..... at least, not in the manuals I have.
afsipmh@cid.aes.doe.CA (Patrick Hertel) (08/20/90)
In article <3440002@hprnd.HP.COM> egt@hprnd.HP.COM (Eric Tausheck) writes: >> > I need to do the following in vi. >> > ccccoooossss -> cos >> >> :%s/cc*/c/g >> :%s/oo*/o/g >> :%s/ss*/s/g >> ---------- > > Wouldn't one of the following be more accurate? >:%s/cc*oo*ss*/cos/g >:%s/ccccoooossss/cos/g > I have seen a lot of answres of this type in the last fwe days but it seems obvious, to me at any rate , that the question reffered to a more general solution. ie he wants to know how to edit out the quadrupling of letters not just the specific case of ccccoooossss. I am not experienced enough with vi to supply an answer but so far I haven't seen it. -- Pat Hertel Canadian Meteorological Centre Analyst/Programmer 2121 N. Service Rd. phertel@cmc.aes.doe.ca Dorval,Quebec Environment Canada CANADA H9P1J3
was@hp-lsd.COS.HP.COM (Bill Stubblebine) (08/20/90)
> First off, if the sources I sent patkar@..... work, this is a dead > issue. the intent was to get rid of the overstriking done by nroff > (etc) for bold type. It would have helped if you had said initially that these were nroff overstrikes. In this case, it's easier to cure the problem than to cure the symptoms. Try: ... | nroff [options] | col -b NAME col - filter reverse line-feeds and backspaces SYNOPSIS col [ -blfxp ] DESCRIPTION (stuff deleted) If the -b option is given, col assumes that the output device in use is not capable of backspacing. In this case, if two or more characters are to appear in the same place, only the last one read will be output. (more stuff deleted) > Second, and here's where I may simply have more to learn about vi (which > is the case all the time.....), with your search/replace commands above, > what happens with strings such as "success", "pool", and so on? Well, you have a good point here. No editor I know of (other than a guy with a green eye shade and a red pencil) knows that two 'o's belong in 'loop', but only one belongs in 'lop'. > I guess the real question should be, does the "*" tell vi that you're > talking about strings of more than 2 of that character? The '*' in a regular expression (RE) means 0 or more occurrences of the character preceding the '*'. Thus, the RE 'cc*' means one or more adjacent instances of the character 'c'. If you cannot locate col(1) at your site, try this shell script, colorfully known as a 'here document': ex << ! r file %s/aaaa/a/g %s/bbbb/b/g ... %s/zzzz/z/g %s/AAAA/A/g %s/BBBB/B/g ... %s/ZZZZ/Z/g (...etc...) w q ! Bill Stubblebine Hewlett-Packard Logic Systems Div. 8245 N. Union Blvd. Colorado Springs, CO 80920 was@hp-lsd.hp.com (Internet) (719) 590-5568
reddy@lion.austin.ibm.com (/150000;Austin) (08/20/90)
Hi, Is there a quick way to make a range of lines in a file to be of same length ? In other words I want the lines in question to contain the same number of characters (padded with blanks if necessary at the end). (awk, sed or shell script solutions are fine too). Satish
fyl@ssc.UUCP (Phil Hughes) (08/21/90)
In article <1990Aug16.130232@ecn.purdue.edu>, patkar@ecn.purdue.edu (Anant Y Patkar) writes: > Just to clear the confusion. What I needed was to change > quadruple occurances of any letter to a single occurance. > That means if I use individual substitute commands for each > letter, I am stuck with 52 commands (some files I have need > more commands to take care of some other characters). > The solutions I > received my mail indicate that (possibly) there is no way to > do this using the "ex" commands. You have to invoke some UNIX > commands like 'sed' from inside vi to do this. Here is my solution which worked fine in ed but does not work in vi (which I don't understand). I expect that ex would be happy with it as well but didn't try it. Anyway, the idea is to just look for any character followed by 3 more of the same character and replace it by one of the character. %s/\(.\)\1\1\1/\1/g (And to think I avoided the "remember this" constructs for years. :-) -- Phil Hughes, SSC, Inc. P.O. Box 55549, Seattle, WA 98155 (206)FOR-UNIX uunet!pilchuck!ssc!fyl or attmail!ssc!fyl (206)527-3385
gwc@root.co.uk (Geoff Clare) (08/22/90)
Somebody asked: >>> > I need to do the following in vi. >>> > ccccoooossss -> cos afsipmh@cid.aes.doe.CA (Patrick Hertel) wrote: > I have seen a lot of answres of this type in the last fwe days but it >seems obvious, to me at any rate , that the question reffered to a more >general solution. ie he wants to know how to edit out the quadrupling >of letters not just the specific case of ccccoooossss. I am not >experienced enough with vi to supply an answer but so far I haven't seen it. That's also how I read the original question. My first attempt was to try this: :s/\(.\)\1\1\1/\1/g which works in "ed", but for some reason it didn't work in "vi". So the next best thing (if you really must do it from within vi) is to use '!' to filter the relevant text through "sed". E.g. for the whole file: 1G!Gsed 's/\(.\)\1\1\1/\1/g' or, if the backspaces from nroff emboldening are still there: 1G!Gsed 's/\(.\)^H\1^H\1^H\1/\1/g' (where ^H is entered using CTRL-V CTRL-H). -- Geoff Clare <gwc@root.co.uk> (Dumb American mailers: ...!uunet!root.co.uk!gwc) UniSoft Limited, Hayne Street, London EC1A 9HH, England. Tel: +44-71-315-6600
dolf@idca.tds.PHILIPS.nl (Dolf Grunbauer) (08/23/90)
In article <2394@root44.co.uk> gwc@root.co.uk (Geoff Clare) writes:
<Somebody asked:
<
<>>> > I need to do the following in vi.
<>>> > ccccoooossss -> cos
<
<[ stuff deleted about how to use sed to deal with it ]
<
<or, if the backspaces from nroff emboldening are still there:
<
< 1G!Gsed 's/\(.\)^H\1^H\1^H\1/\1/g'
<
<(where ^H is entered using CTRL-V CTRL-H).
This can be done in plain vi by typing:
%s/^H.//g
As the backspace sequence is used only to reprint the preceding character and
with this command all duplications are removed.
--
Dolf Grunbauer Tel: +31 55 433233 Internet dolf@idca.tds.philips.nl
Philips Information Systems UUCP ...!mcsun!philapd!dolf
Some kind of happiness is measured out in milesgbastin@x102c.harris-atd.com (Gary Bastin 60293) (08/23/90)
In article <1117@ssp11.idca.tds.philips.nl> dolf@idca.tds.PHILIPS.nl (Dolf Grunbauer) writes: >In article <2394@root44.co.uk> gwc@root.co.uk (Geoff Clare) writes: ><Somebody asked: >< ><>>> > I need to do the following in vi. ><>>> > ccccoooossss -> cos >< >This can be done in plain vi by typing: > > %s/^H.//g > >As the backspace sequence is used only to reprint the preceding character and >with this command all duplications are removed. Nope. Afraid this one won't work, either. It DOES clean up all the occurences of "^H." But there is a subtle problem. For example, on my system, the man pages for csh when piped to a file create lots of "^H." occurences, which this does clean. This works fine for overstrikes used to just embolden particular words. But also generated are lines of the form _^HC_^Hs_^Hh, which generates "Csh" with an underline. Using the global replacement of "^H." with nothing completely obliterates the stuff that uses the overstrike for really making underlined words, leaving just the underline characters with nothing above them! Which is not what is desired, I am sure. (Not a flame, just a subtle point!) Gary Bastin /-/-/ Internet: gbastin@x102c.ess.harris.com Mail Stop 102-4826 | phone: (407) 729-3045 Harris Corporation GASD | packet: WB4YAF @ N4JLR.FL.USA.NA P.O.B. 94000, Melbourne FL 32902 Speaking from, but not for, Harris!
new@ee.udel.edu (Darren New) (08/23/90)
>> %s/^H.//g >>As the backspace sequence is used only to reprint the preceding character and >>with this command all duplications are removed. Much better is %s/.^H//g which throws out the character *before* the backspace in case it is an underline. You know this is right because the backspace when displayed overwrites the *previous* character, not the next. -- Darren -- --- Darren New --- Grad Student --- CIS --- Univ. of Delaware ---
peter@ficc.ferranti.com (Peter da Silva) (08/24/90)
In article <1117@ssp11.idca.tds.philips.nl> dolf@idca.tds.PHILIPS.nl (Dolf Grunbauer) writes: > %s/^H.//g This will be fun if you're underlining. How about: %s/.^H//g Which more accurately emulates what actually happens, anyway. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
egt@hprnd.HP.COM (Eric Tausheck) (08/24/90)
> > I had the need to do this kind of editing too until I learned about > col -b. > > If what you started with had a bunch of back-spaces between the > letters too (as in an nroff or man output), the filter "col -b" does a > nice job of getting rid of all this junk for you. I should have added that before I discovered the "col -b" filter, I used to execute the following vi command to nroff output: :%s/.^H//g (where the "^H" is entered with <ctrl>V<ctrl>H) On a more interesting note... > > :s/\(.\)\1\1\1/\1/g > >which works in "ed", but for some reason it didn't work in "vi". So the >next best thing (if you really must do it from within vi) is to use '!' to >filter the relevant text through "sed". E.g. for the whole file: > > 1G!Gsed 's/\(.\)\1\1\1/\1/g' This is very powerful (of ed and sed). Thanks for sharing this gem regarding sed/ed. A shame vi/ex doesn't implement it because it would (almost?) give vi/ex true complete regular expression matching capabilities. I'm very rusty on my formal language theory so for all I know this (ability to self reference to patterns during regular expression matching) may have given vi/ex full regular expression matching capabilities - anyone out there care to speculate? Eric Tausheck email: egt@hprnd.hp.com Hewlett Packard Co.
roland@ai.mit.edu (Roland McGrath) (08/24/90)
No, don't replace ^H. with nothing; replace .^H with nothing to get rid of overstriking and underlining. -- Roland McGrath Free Software Foundation, Inc. roland@ai.mit.edu, uunet!ai.mit.edu!roland