replacing text - perl, awk....

[hpuxsa]

Hi,
I have many entries like below in a file
/disk1/oracle/HP/hp01.dbf
to
/disk1/oracle/HP/hp01.dbf

/disk2/oracle/HP/hp02.dbf
to
/disk2/oracle/HP/hp02.dbf

I would like to change them as follows

/disk1/oracle/HP/hp01.dbf
to
/disk1_dev/oracle_dev/hp01.dbf

/disk2/oracle/HP/hp02.dbf
to
/disk2_dev/oracle_dev/HP/hp02.dbf

Regards,
Steven

[Francisco J. Soler · Accepted solution]

Hi Steven,

my solution,

Create a script in awk like this:

----- s.awk -------------------
BEGIN {
flag_to=0
FS="/"
OFS="/"
}

/to/ { flag_to=1 }
/disk1/ { if (flag_to==1) $2="disk1_dev" ; flag_to=0}
/disk2/ { if (flag_to==1) $2="disk1_dev" ; flag_to=0}
{ print }
--------------------------------

Then run it:

awk -f s.awk filein > fileout

Hope this helps.

Frank.

[H.Merijn Brand (procura]

# perl -pi -e's{\b(disk\d|oracle)/}{$1_dev/};s{/HP/}{/}' your_file

Enjoy, Have FUN! H.Merijn

[Thierry Poels_1]

hi,

these are the datafiles of an Oracle DB. You probably got this list with an SQL statement. How about modifying this SQL statement and "replacing" the second part.

regards,
Thierry.

[Francisco J. Soler]

Hi Steven,

I have made a mistake, i think you already have seen it.

The statement with /disk2/ is not:
/disk2/ { if (flag_to==1) $2="disk1_dev" ; flag_to=0}

must be:
/disk2/ { if (flag_to==1) $2="disk2_dev" ; flag_to=0}

Frank

[Hein van den Heuvel]

So do you want that HP always gone, or just sometimes. If always:

$ perl -pe 's;/oracle/HP;_dev/oracle_dev;' old > new


Hein.

[Abdul Rahiman]

IF HP always exist,
sed "s/\/oracle\//\_dev\/oracle_dev\//g" file.txt

regds,
Abdul.