use multiple field seperators with awk

[renarios]

Hi all,

In a testcase, my /etc/oratab looks like this:
# Oracle 8 stuff (ocho)
tst1:/mnt/u01/app/oracle/product/8.1.7:Y
# Oracle 9 stuff (el nino)
tst2:/mnt/u01/app/oracle/product/9.2.0:Y
tst3:/mnt/u01/app/oracle/product/9.2.0:N

for a backup script I need all "active" Oracle 9i databases. I came up with this:
DBLIST=$(awk -F: '$1 !~ /^#/ && $3 ~ /Y$/ {print $1}' ${ORATAB})
, but that does only filter the "Y" and comment lines. How can I filter the "9" (and 10 later on) out of the lines?
Best would be use something with >8

I'm struggling with split, but I'm stuck
Any ideas?

Cheerio,

Renarios

[Sudeesh]

Hello,

I just did what you have said....

[rx260-11]/sudhi >cat kk
# Oracle 8 stuff (ocho)
tst1:/mnt/u01/app/oracle/product/8.1.7:Y
# Oracle 9 stuff (el nino)
tst2:/mnt/u01/app/oracle/product/9.2.0:Y
tst3:/mnt/u01/app/oracle/product/9.2.0:N

[rx260-11]/sudhi >awk -F: '$1 !~ /^#/ && $3 ~ /Y$/ {print $1}' kk
tst1
tst2
[rx260-11]/sudhi >


and see only test1,test2 in the o/p

what you want to achive? am I missing something in your question?


Sudeesh

[renarios]

Hi Sudeesh,

Thanks for the quick response.
The output should be only tst2, because tst1 is 8.1.7 and I only want the active 9.2.0 (or in future higher) databases.

thanks,

RenÃ

[renarios]

Latest code is
awk -F: '$1 !~ /^#/ && $3 ~ /Y$/ && split($0, arr, "/") {if ( arr[7] > 8 ) print arr[7] "*" $1 " " $2 " " $3}' ${ORATAB}
, but it does not work.

Has anyone other/ideas ideas?

Cheers,

Renarios

[renarios]

I found a solution:
awk -F: '$1 !~ /^#/ && $3 ~ /Y$/ && split($0, arr, "/") && split(arr[7], oraver, ".") {if ( oraver[1] >= 9 ) print $1}' ${ORATAB}

10 point for Renarios?

Only thing is: How do I get this readable?
If I use BEGIN and END, it doesn't work anymore.
Anyone ideas?

Cheerio,

Renarios

[Alex Lavrov.]

Dont' make it readable, if it works, leave it as is and make a really good comment how it works.

Because when you try to rewrite things so they'll look simplier, you can put there some bugs.

Good comment on this line will be enough I think :)

[curt larson_1]

something a little more readable

awk -F: '$1 !~ /^#/ && $3 ~ /Y$/ {
split($0, arr, "/");
split(arr[7], oraver, ".");
if ( oraver[1] >= 9 ) print $1;
}' ${ORATAB}

[renarios]

Thanks guys,

That should do the trick.
Time for weekend. Have one on me.

Solution:
DBLIST=$( awk -F: '$1 !~ /^#/ && $3 ~ /Y$/ {
split($0, arr, "/");
split(arr[7], oraver, ".");
if ( oraver[1] >= 9 ) print $1;
}' ${ORATAB} )

Cheerio,

Renarios