Here is a quick, non optomized perl script:
while () {
next unless (/^HCA/);
print $_;
while () {
last unless (/^I[XY]\d\d\d\d\d/);
print $_;
}
}
There are more efficient ways to do this then the above perl, but that is
sufficient...
- Marc
On Wed, 9 Feb 2000, Matt Tilley wrote:
Sorry for the long post, but tried to include enough info so that people can
help.
Also, there is probably a better group to post to, but a quick search for
"script" in newsgroups didn't turn up anything that led me to believe that the
question would be better directed to another group (if I'm wrong - please just
point me in the right direction, I'll post there and sorry for the bother).
OK, here goes. My problem stems from dealing with a file that contains
lines. Some of the lines I want to keep, while I want to discard others
(based on the value of the field of the first column). It seems that the
nutshell book deals more with files than with variables (this could just be a
stumbling block of mine)
I am looking for groups of lines that belong together - for instance, whenever
a line starts with "HCA" I want to save that line to a file and append the
following lines as long as they begin with IX? or IY? (the "?" are
place markers - I know that I will have a five digit number there). As soon
as the next HCA comes along, I want to do the same thing again.
Again, there are some lines that contain starting characters different than
those above (or just blank lines) that I just want to discard.
Here is a quick example of the file that I want to work with:
START:
Garbage line that I want to discard
More garbage
HCA would like to keep this line
IY45678 - and keep this line under the one above it (once I get the structure
I can do more)
IX89374 - with the rest.
More garbage that isn't wanted
HCA would like to also keep this line, appending it to those above
IX84956 - but preferably separated by a line
IY74658 - this should be simple but I'm having a mental block. Once over the
block
IY34567 - I feel like a mental switch will close and I will begin to be more
productive
IX84755 - with my scripts (otherwise they seem way to simple (almost like a
batch file)
END!
BTW, I would like to use ksh for this (will look into other ways (such as
perl) after I close my knowledge gap in places like this). My thoughts on
this were this (just to at least show that I've been thinking about it):
cat myfile | while read line
do
for i in $line
do
if [ [ $i = IX? || IY? || HCA ] ]
# would like to exit if true to the while loop and dump the line
# to a file - possibly setting a flag to indicate that HCA is true
(i.e., hcaflag)
# as soon as another HCA or an unwanted word is encountered
# print a blank line, discard unwanted line or start process over
# again
then exit to while loop if true
else discard line and get next line from myfile
fi
if [ hcaflag ]
then if [ [ $i = IX? || IY? ] ]
then echo line to file, etc.
Is this the right way to go about this (or haven't I explained it in enough
detail to understand my problem?).
I'm not in scripting school (although I've been through one a while back), and
I'm not doing this for anyone else but me (just to help me with my scripting
skills). I'm willing to put some time into this, but feel as if there maybe
an easier way that I'm overlooking (am I making this way too difficult?). The
purpose behind this is to expedite the handling of large files indicating the
fileset status on my AIX boxes (yes, I know that this isn't an AIX list, but
I'm always monitoring this list because of my interest in Linux (have one
P-166 running RH6.0 and on/off luck with running RH on my Dec Alpha (both at
home)).
Sorry for the long message, but wanted to be sure that I gave enough
information (hoping that I achieved this).
TIA
- Matt
**
To unsubscribe from this list, send mail to
[EMAIL PROTECTED] with the following text in the
*body* (*not* the subject line) of the letter:
unsubscribe gnhlug
**
**
To unsubscribe from this list, send mail to
[EMAIL PROTECTED] with the following text in the
*body* (*not* the subject line) of the letter:
unsubscribe gnhlug
**
