I've been learning sed and whilst the complicated things seem to work,
some of the simple ones don't.  For example:

I have a file with records of the form:
PMID- 14974909
OWN - NLM
STAT- completed
DA  - 20040225
DCOM- 20040325
IS  - 0300-0664
VI  - 59
IP  - 6
DP  - 2003 Dec
PG  - 690-8
FAU - Nugent, Ailish G
AU  - Nugent AG
FAU - Leung, Kin-Chuen
AU  - Leung KC
FAU - Sullivan, David
AU  - Sullivan D
FAU - Reutens, Anne T
AU  - Reutens AT
FAU - Ho, Ken K Y
AU  - Ho KK
LA  - eng
PT  - Clinical Trial
PT  - Journal Article
PT  - Randomized Controlled Trial
PL  - England
TA  - Clin Endocrinol (Oxf)
JID - 0346653
RN  - 0 (Carrier Proteins)
SO  - Clin Endocrinol (Oxf) 2003 Dec;59(6):690-8.

How can I get sed (or grep), to output only the lines tagged with,
say, AU and PT (in the same order as in the source file).  I though '()'
allowed me to group patterns a la:

cat filename |grep (^AU,^PT)

but, apparently not.
I thought that maybe:

cat filename |sed /^[^PT,^AU]/\!d
might do it, but no.

Also how can I get sed to remove CR/LFs at the end of specific lines?

try this:
cat filename | grep -e ^AU -e PT 

On Fri, 02 Jul 2004 11:37:35 +0200, Ed wrote:

correction:
cat filename | grep -e ^AU -e ^PT

On Fri, 02 Jul 2004 13:29:36 +0200, John Stolz <china-rider@wanadoo.fr> wrote:
| I've been learning sed and whilst the complicated things seem to work,
| some of the simple ones don't.  For example:
|
| I have a file with records of the form:
| PMID- 14974909
| OWN - NLM
| STAT- completed
| DA  - 20040225
| DCOM- 20040325
| IS  - 0300-0664
| VI  - 59
| IP  - 6
| DP  - 2003 Dec
| PG  - 690-8
| FAU - Nugent, Ailish G
| AU  - Nugent AG
| FAU - Leung, Kin-Chuen
| AU  - Leung KC
| FAU - Sullivan, David
| AU  - Sullivan D
| FAU - Reutens, Anne T
| AU  - Reutens AT
| FAU - Ho, Ken K Y
| AU  - Ho KK
| LA  - eng
| PT  - Clinical Trial
| PT  - Journal Article
| PT  - Randomized Controlled Trial
| PL  - England
| TA  - Clin Endocrinol (Oxf)
| JID - 0346653
| RN  - 0 (Carrier Proteins)
| SO  - Clin Endocrinol (Oxf) 2003 Dec;59(6):690-8.
|
| How can I get sed (or grep), to output only the lines tagged with,
| say, AU and PT (in the same order as in the source file).  I though '()'
| allowed me to group patterns a la:
|
| cat filename |grep (^AU,^PT)


Almost there. Try:

  grep -E '^AU|^PT' filename

also 'egrep' is shorthand for 'grep -E'

The pipe '|' means 'or', and the bracket do group the expression. You need the
quotes to stop the shell seeing the '|' character.

You don't need the grouping for this simple expression, but you could do:

  grep -E '^(AU|PT)' filename


| but, apparently not.
| I thought that maybe:
|
| cat filename |sed /^[^PT,^AU]/\!d
| might do it, but no.
|
| Also how can I get sed to remove CR/LFs at the end of specific lines?


Not sure how to do this is sed, sorry.


-- 
Reverend Paul Colquhoun, ULC.    http://andor.dropbear.id.au/~paulcol
     Asking for technical help in newsgroups?  Read this first:
        http://catb.org/~esr/faqs/smart-questions.html#intro

> | PMID- 14974909
> | OWN - NLM
> | STAT- completed
> | DA  - 20040225
> | DCOM- 20040325
> | IS  - 0300-0664
> | VI  - 59
> | IP  - 6
> | DP  - 2003 Dec
> | PG  - 690-8
> | FAU - Nugent, Ailish G
> | AU  - Nugent AG
> | FAU - Leung, Kin-Chuen
> | AU  - Leung KC
> | FAU - Sullivan, David
> | AU  - Sullivan D
> | FAU - Reutens, Anne T
> | AU  - Reutens AT
> | FAU - Ho, Ken K Y
> | AU  - Ho KK
> | LA  - eng
> | PT  - Clinical Trial
> | PT  - Journal Article
> | PT  - Randomized Controlled Trial
> | PL  - England
> | TA  - Clin Endocrinol (Oxf)
> | JID - 0346653
> | RN  - 0 (Carrier Proteins)
> | SO  - Clin Endocrinol (Oxf) 2003 Dec;59(6):690-8.
> |
> | How can I get sed (or grep), to output only the lines tagged with,
> | say, AU and PT (in the same order as in the source file).  I though '()'
> | allowed me to group patterns a la:
> |
> | cat filename |grep (^AU,^PT)
>
>
> Almost there. Try:
>
>   grep -E '^AU|^PT' filename

I like the grep solution but
cat filename | grep -E '^AU  -|^PT  -'
is more appropriate unless you know more about the tags than I do.

Regards...Dan.

On 2004-07-02, Dan Skinner wrote:
>
> cat filename | grep -E '^AU  -|^PT  -'
> is more appropriate unless you know more about the tags than I do.

    There's no need for cat:

grep -E '^AU  -|^PT  -' filename

-- 
    Chris F.A. Johnson                  http://cfaj.freeshell.org/shell
    ===================================================================
    My code (if any) in this post is copyright 2004, Chris F.A. Johnson
    and may be copied under the terms of the GNU General Public License

On Fri, 02 Jul 2004 11:38:21 +0200, Ed hath writ:
> On Fri, 02 Jul 2004 11:37:35 +0200, Ed wrote:
>
> correction:
> cat filename | grep -e ^AU -e ^PT

  *Correct* correction:

  grep -e ^AU -e ^PT filename

Jonesy

On Fri, 02 Jul 2004 18:02:10 +0000, Chris F.A. Johnson wrote:

> On 2004-07-02, Dan Skinner wrote:
>>
>> cat filename | grep -E '^AU  -|^PT  -'
>> is more appropriate unless you know more about the tags than I do.
> 
>     There's no need for cat:
> 
> grep -E '^AU  -|^PT  -' filename

Thanks everyone - this works a treat now.