Regexp script needed, please help

Mr. Alex · 05-21-2013, 05:41 PM

Hello people.
Help me to make a simple regexp script please. There's a csv-file with lines of text. I need script to find all "new line" brakes in this text file and those "new line" brakes that are not followed by at least three digits in a row, change to " ". I guess the script is one line and very simple, I just don't have time and real need to learn sed regexp or any other regexps.

unSpawn · 05-21-2013, 06:08 PM

Quote:

Originally Posted by Mr. Alex

Help me to make a simple regexp script please.

Do make an effort to post anything we can actually help you with I'd say.

danielbmartin · 05-21-2013, 06:25 PM

Quote:

Originally Posted by Mr. Alex

Hello people.
Help me to make a simple regexp script please. There's a csv-file with lines of text. I need script to find all "new line" brakes in this text file and those "new line" brakes that are not followed by at least three digits in a row, change to " ". I guess the script is one line and very simple, I just don't have time and real need to learn sed regexp or any other regexps.

Help us to help you. Provide a sample input file (10-15 lines will do). Construct a sample output file which corresponds to your sample input and post both samples here. With "Before and After" examples we can better understand your needs and also judge if our proposed solution fills those needs.

Daniel B. Martin

grail · 05-22-2013, 01:41 AM

I am with the above ... without an example it would prove difficult to help.

Also, you mention , which would seem to imply passing html, if so this would be useful information.

pan64 · 05-22-2013, 03:06 AM

not to speak about that you can find simple csv regexp on the net: http://dotnetslackers.com/Regex/re-7...commas_yo.aspx

Mr. Alex · 05-23-2013, 09:22 AM

Quote:

Also, you mention , which would seem to imply passing html, if so this would be useful information.

It is for saving as HTML, but the source files are csv.

Quote:

Originally Posted by danielbmartin

Help us to help you. Provide a sample input file (10-15 lines will do). Construct a sample output file which corresponds to your sample input and post both samples here. With "Before and After" examples we can better understand your needs and also judge if our proposed solution fills those needs.

Daniel B. Martin

Code:

=== before ===

1234;aaa;bbb;ccc;ddd
6352;jncasdu;qweyghq;cnhjab;
234;ajbsndc;acyue;aciusdh;


ksjdhcaiuh,

asdnclius
3478;sadcbh;wdoehuiq;csanld
1290;asdhuoci;qdbhejkw;

1. cshdb
2. huscdio
3. csdhuioa
4. wduqhie
2763;chsduaio;wdebhk;scdhuao

=== after ===

1234;aaa;bbb;ccc;ddd
6352;jncasdu;qweyghq;cnhjab;
234;ajbsndc;acyue;aciusdh;<br><br><br>ksjdhcaiuh,<br><br>asdnclius
3478;sadcbh;wdoehuiq;csanld
1290;asdhuoci;qdbhejkw;<br><br>1. cshdb<br>2. huscdio<br>3. csdhuioa<br>4. wduqhie
2763;chsduaio;wdebhk;scdhuao

Quote:

Originally Posted by pan64

not to speak about that you can find simple csv regexp on the net: http://dotnetslackers.com/Regex/re-7...commas_yo.aspx

I don't need to split csv file.

danielbmartin · 05-23-2013, 10:10 AM

With this InFile ...

Code:

1234;aaa;bbb;ccc;ddd
6352;jncasdu;qweyghq;cnhjab;
234;ajbsndc;acyue;aciusdh;


ksjdhcaiuh,

asdnclius
3478;sadcbh;wdoehuiq;csanld
1290;asdhuoci;qdbhejkw;

1. cshdb
2. huscdio
3. csdhuioa
4. wduqhie
2763;chsduaio;wdebhk;scdhuao

... this brute-force code ...

Code:

 sed 's/\(^[0-9][0-9][0-9]\)/~\1/' $InFile  \
|paste -sd'%' -   \
|sed 's/^~//'     \
|sed 's/%~/\n/g'  \
|sed 's/%/<br>/g' \
> $OutFile

... produced this OutFile ...

Code:

1234;aaa;bbb;ccc;ddd
6352;jncasdu;qweyghq;cnhjab;
234;ajbsndc;acyue;aciusdh;<br><br><br>ksjdhcaiuh,<br><br>asdnclius
3478;sadcbh;wdoehuiq;csanld
1290;asdhuoci;qdbhejkw;<br><br>1. cshdb<br>2. huscdio<br>3. csdhuioa<br>4. wduqhie
2763;chsduaio;wdebhk;scdhuao

Daniel B. Martin

grail · 05-23-2013, 10:22 AM

Maybe something like:

Code:

awk 'NR > 1{ORS = /^[0-9]+;/?RT:"<br>"; print x}{x=$0}END{print x}' file

danielbmartin · 05-23-2013, 11:48 AM

Quote:

Originally Posted by grail

Maybe something like:

Code:

awk 'NR > 1{ORS = /^[0-9]+;/?RT:"<br>"; print x}{x=$0}END{print x}' file

Short and sweet!

Minor nitpick:

Quote:

I need script to find all "new line" brakes in this text file and those "new line" brakes that are not followed by at least three digits in a row, change to " ".

Your awk doesn't handle the three-consecutive-digits test.

Daniel B. Martin

grail · 05-23-2013, 12:39 PM

Fair enough ... change + to {3,}

danielbmartin · 05-23-2013, 12:56 PM

Quote:

Originally Posted by grail

Fair enough ... change + to {3,}

With this InFile ...

Code:

1234;aaa;bbb;ccc;ddd
6352;jncasdu;qweyghq;cnhjab;
23;ajbsndc;acyue;aciusdh;


ksjdhcaiuh,

asdnclius
3478;sadcbh;wdoehuiq;csanld
1290;asdhuoci;qdbhejkw;

1. cshdb
2. huscdio
3. csdhuioa
4. wduqhie
2763;chsduaio;wdebhk;scdhuao

... this awk ...

Code:

awk 'NR > 1{ORS = /^[0-9]{3,};/?RT:"<br>"; print x}
  {x=$0}END{print x}' $InFile >$OutFile

... produced this OutFile ...

Code:

1234;aaa;bbb;ccc;ddd<br>6352;jncasdu;qweyghq;cnhjab;<br>23;ajbsndc;acyue;aciusdh;<br><br><br>ksjdhcaiuh,<br><br>asdnclius<br>3478;sadcbh;wdoehuiq;csanld<br>1290;asdhuoci;qdbhejkw;<br><br>1. cshdb<br>2. huscdio<br>3. csdhuioa<br>4. wduqhie<br>2763;chsduaio;wdebhk;scdhuao<br>

... which looks wrong.

Daniel B. Martin

grail · 05-23-2013, 02:01 PM

And I would be guessing you are using a pre version 4 of gawk? Otherwise it works fine

danielbmartin · 05-23-2013, 04:28 PM

Quote:

Originally Posted by grail

And I would be guessing you are using a pre version 4 of gawk?

Ahh, yes.

Code:

daniel@daniel-desktop:~$ gawk --version
GNU Awk 3.1.6
Copyright (C) 1989, 1991-2007 Free Software Foundation.

Daniel B. Martin

grail · 05-24-2013, 03:04 AM

I haven't had to try it in a while but there is a --re-interval (from memory) which I think may help.

danielbmartin · 05-24-2013, 08:05 AM

Quote:

Originally Posted by grail

... there is a --re-interval ... which I think may help.

Yes!

This works with GNU awk 3.1.6 ...

Code:

awk --re-interval 'NR > 1{ORS = /^[0-9]{3,};/?RT:"<br>"; print x}
  {x=$0}END{print x}' $InFile >$OutFile

Daniel B. Martin