Search Results

Like that?


bash-3.2$ awk 'BEGIN{x=1} /DEFINE SCHEMA/,/\)\;/ {if (NF==2) { $1="t"x" as filler"; print; x++ }}' filename
t1 as filler VARCHAR(10)

awk '{print "t"++i" as filler VARCHAR"}' extract_file

Using awk


awk '{print substr($0,length($0)-7)}' filename

Try:
for i in *.tpt.ctl
do
var=${i%.tpt.ctl}
printf "%s\n" "$var"
done

Or using sed:

sed 's/\(.*\).\{8\}$/\1/' file

@krishmaths
Simpler approach
awk -F. '{print $1}'

But why do it complicate, when we can do it the Scrutinizer way :)

echo abc.tpt.ctl | rev | cut -c 9- | rev
abc

bash-3.2$ cat x
abc.tpt.ctl
bdc.tpt.ctl
cdw.tpt.ctl
bash-3.2$ cat x | rev | cut -c 9- | rev
abc
bdc
cdw

Having real life data helps allot, why did you not post it before :(
Now script only looks for end ) after pattern and ( is found.

awk '/APPLY/ {t=1} f && /\)/{exit} f; /\(/ && t {f=1}' file
...

@Jotne, Thanks for noting but my solution is not generic and applies to the input data posted in this thread. I hope we both have given enough directions now :)

If you read from a file and write to it at the same time, all you get is a corrupted file.
sed has options for in-file editing --> the -i switch

sed -i.bak 's/bteq/tpt/g' filelist.txt

Now you...

awk '!(NR%2){$(NF+1)=";"}1' filename

awk '!(NR%2){$NF=$NF";"}1' infile

--ahamed

Just check how many lines of your message has header files.

Assume it first two lines are header file here.. use

"psql -c "\dt" | awk 'NR>2{print$3}'"

From what we can see in the desired output, it should start with line 4, so pamu's proposal should be slightly modified to using NR>3.

I think that if you experiment a little with the printf statements you should be able to do this yourself. What have you tried so far?