Hi Phil, Very easy, as you said. It worked fine as expected. Look at this: source data: INITIAL 1000 AAA INITIAL 2000 AAA INITIAL 3000 AAA INITIAL 4000 BBB the script: #!/bin/sh cat lista | while read arq do echo $arq | awk '{ sub(/INITIAL [0-9]/,"INITIAL 8124"); print }' done after running the script the result is: INITIAL 8124000 AAA INITIAL 8124000 AAA INITIAL 8124000 AAA INITIAL 8124000 BBB Note: In order to get the purpose I need to run again like that: #!/bin/sh cat lista | while read arq do echo $arq | awk '{ sub(/INITIAL 8124000/,"INITIAL 8124"); print }' done Issue solved. Thanks for help. Best Regards Eriovaldo On Sun, Jul 10, 2011 at 9:30 AM, Phillip Jones <phil@xxxxxxxxxx> wrote: > Hi, > > It's easy with sed: > > sed 's/INITIAL [0-9]*/INITIAL 8124/' > > Thanks, > > Phil > > On Sun, Jul 10, 2011 at 12:55 PM, Eriovaldo Andrietta < > ecandrietta@xxxxxxxxx> wrote: > >> Hi friends, >> >> I generated a file using imp / indexfile in order to know objects inside >> the .dmp. >> Then I saw that the initial value is too big to import it. >> >> How can I do an awk code in order to change the folloing string : >> >> from INITIAL 52428800 FREELISTS >> to INITIAL 8124 FREELISTS >> >> Note: in the indexfile has a lot of initial values different, so the awk >> must consider different values. >> >> I have this one, but I don´t know how to get and change the old value >> (the next word after the word "INITIAL". >> >> Look: >> >> cat lista1 | awk '{ sub(/INITIAL/, "INITIAL 8124"); print }' >> >> Best Regards >> Eriovaldo >> >> >> >