awk question
in [Shell]
|
From: Laurianne Gardeux on 9 May 2010 17:59 Rocky Jr. à écrit : > On May 10, 1:55 am, Janis Papanagnou <janis_papanag...(a)hotmail.com> > wrote: >> Janis Papanagnou wrote: >> > Rocky Jr. wrote: >> >> My data file >> >> ---- >> >> zone "data1.txt" { >> >> type gb2; >> >> file "/nas/data1.txt"; >> >> }; >> >> >> zone "delta.doc" { >> >> type gb2; >> >> file "/nas/delta.doc"; >> >> }; >> >> >> zone "meta.exe" { >> >> type gb2; >> >> file "/nas/meta.exe"; >> >> }; >> >> ---- >> >> >> I need to select text between zone and }; so I wrote awk '/^zone/, >> >> /^};$/' input.file >> >> >> This works great and but I need to store selected data in file >> >> called delta.doc or meta.exe, in short select the following and >> >> store to data1.txt: >> >> >> zone "data1.txt" { >> >> type gb2; >> >> file "/nas/data1.txt"; >> >> }; >> >> >> The file has over 100s of entry and I need to create 100 new files >> >> from a single file. How do I ask awk to select data1.txt from 'zone >> >> "data1.txt"' ? >> >> > I suppose it would be as good for your task to not only select a >> > single block but create the files in one pass? Then try... >> >> > awk 'BEGIN { c=1 ; f="f"c } >> > /^zone/, /^};$/ { print > f } >> > !NF { close(f) ; f="f"++c }' >> >> > which will create files f1..f100 from the respective blocks. >> >> And another (shorter) solution... >> >> awk 'BEGIN { RS="" } /^zone/, /^};$/ { f="f"++c ; print > f ; >> close(f) }' >> >> >> >> >> >> > Janis > > Thanks Janis! > > It is working as {f1..f100}. Is there any way we can replace f1 with > actual name such as data1.txt ans so on? awk 'BEGIN { RS="" }/^zone/, /^};$/ {split($0, a, "\""); print > a[2] }' LG
From: John W. Krahn on 9 May 2010 19:02 Rocky Jr. wrote: > My data file > ---- > zone "data1.txt" { > type gb2; > file "/nas/data1.txt"; > }; > > zone "delta.doc" { > type gb2; > file "/nas/delta.doc"; > }; > > zone "meta.exe" { > type gb2; > file "/nas/meta.exe"; > }; > ---- > > I need to select text between zone and }; so I wrote > awk '/^zone/, /^};$/' input.file > > This works great and but I need to store selected data in file called > delta.doc or meta.exe, in short select the following and store to > data1.txt: > > zone "data1.txt" { > type gb2; > file "/nas/data1.txt"; > }; > > The file has over 100s of entry and I need to create 100 new files > from a single file. How do I ask awk to select data1.txt from 'zone > "data1.txt"' ? perl -pe'BEGIN{ $/ = "};\n" } /zone "([^"]+)"/ and open STDOUT, ">>", $1' input.file John -- The programmer is fighting against the two most destructive forces in the universe: entropy and human stupidity. -- Damian Conway
From: Janis Papanagnou on 10 May 2010 02:25 Laurianne Gardeux wrote: > Rocky Jr. à écrit : > >> On May 10, 1:55 am, Janis Papanagnou <janis_papanag...(a)hotmail.com> >> wrote: >>> Janis Papanagnou wrote: >>>> Rocky Jr. wrote: >>>>> My data file >>>>> ---- >>>>> zone "data1.txt" { >>>>> type gb2; >>>>> file "/nas/data1.txt"; >>>>> }; >>>>> zone "delta.doc" { >>>>> type gb2; >>>>> file "/nas/delta.doc"; >>>>> }; >>>>> zone "meta.exe" { >>>>> type gb2; >>>>> file "/nas/meta.exe"; >>>>> }; >>>>> ---- >>>>> I need to select text between zone and }; so I wrote awk '/^zone/, >>>>> /^};$/' input.file >>>>> This works great and but I need to store selected data in file >>>>> called delta.doc or meta.exe, in short select the following and >>>>> store to data1.txt: >>>>> zone "data1.txt" { >>>>> type gb2; >>>>> file "/nas/data1.txt"; >>>>> }; >>>>> The file has over 100s of entry and I need to create 100 new files >>>>> from a single file. How do I ask awk to select data1.txt from 'zone >>>>> "data1.txt"' ? >>>> I suppose it would be as good for your task to not only select a >>>> single block but create the files in one pass? Then try... >>>> awk 'BEGIN { c=1 ; f="f"c } >>>> /^zone/, /^};$/ { print > f } >>>> !NF { close(f) ; f="f"++c }' >>>> which will create files f1..f100 from the respective blocks. >>> And another (shorter) solution... >>> >>> awk 'BEGIN { RS="" } /^zone/, /^};$/ { f="f"++c ; print > f ; >>> close(f) }' >>> >>> >>> >>> >>> >>>> Janis >> Thanks Janis! >> >> It is working as {f1..f100}. Is there any way we can replace f1 with >> actual name such as data1.txt ans so on? In case I'd above have misunderstood the actual filename the OP actually wants, please note that... > > > awk 'BEGIN { RS="" }/^zone/, /^};$/ {split($0, a, "\""); print > a[2] }' ....it's better to close() the unused open file handles, and you can also simplify the code using the builtin field splitting... awk 'BEGIN { RS="" ; FS="\"" } /^zone/, /^};$/ { print > $2 ; close($2) }' Janis > > LG
From: Laurianne Gardeux on 10 May 2010 06:38 Janis Papanagnou à écrit : > Laurianne Gardeux wrote: >> awk 'BEGIN { RS="" }/^zone/, /^};$/ {split($0, a, "\""); print > a[2] >> }' > > ...it's better to close() the unused open file handles, and you can also > simplify the code using the builtin field splitting... > > awk 'BEGIN { RS="" ; FS="\"" } /^zone/, /^};$/ { print > $2 ; close($2) > }' > > > Janis Oh yes, thank you ! LG
First
|
Prev
|
Pages: 1 2 Prev: Exit status of process by PID Next: Running a script through Automatic scheduler |