|
|
|
|
|
|
| Author |
Message |
Rakesh Sharma
*nix forums beginner
Joined: 28 Feb 2005
Posts: 36
|
 Posted: Fri Jul 21, 2006 10:07 am Post subject:
Re: joining lines in ksh
|
|
|
jctown@nb.sympatico.ca wrote:
| Quote: | I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
...............[snip] |
sed -ne '/^category/h;//!G;s/\(.*\)\n\(.*\)/\2,\1/p' inputfile |
|
| Back to top |
|
 |
jctown@nb.sympatico.ca
*nix forums beginner
Joined: 19 Jul 2006
Posts: 6
|
| Posted: Thu Jul 20, 2006 1:17 am Post subject:
Re: joining lines in ksh
|
|
|
Chris F.A. Johnson wrote:
| Quote: | On 2006-07-19, jctown@nb.sympatico.ca wrote:
[snip]
Here's my version of your script... my posix complained about [ $? -eq
0 ] so I put it back to [[ ]]
What, exactly, was the complaint? [[...]] is not POSIX.
count=1
while read line junk ## $junk will contain everything after the first word
do
case $line in
*".jar"*) true ;;
*) false ;;
esac
if [[ $? -eq 0 ]]
then
save_category=$line
else
echo $count "," $save_category "," $line
let count=$count+1
As I said before, that is not POSIX (actually, I think I said it
was not portable -- same thing).
count=$(( $count + 1 ))
fi
done < xxx.txt > new_one.txt
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
|
I put a she-bang at the top to force ksh....and it worked with [..] No
syntax errors |
|
| Back to top |
|
|
Chris F.A. Johnson
*nix forums Guru
Joined: 20 Feb 2005
Posts: 2268
|
| Posted: Thu Jul 20, 2006 12:45 am Post subject:
Re: joining lines in ksh
|
|
|
On 2006-07-19, jctown@nb.sympatico.ca wrote:
[snip]
| Quote: |
Here's my version of your script... my posix complained about [ $? -eq
0 ] so I put it back to [[ ]]
|
What, exactly, was the complaint? [[...]] is not POSIX.
| Quote: | count=1
while read line junk ## $junk will contain everything after the first word
do
case $line in
*".jar"*) true ;;
*) false ;;
esac
if [[ $? -eq 0 ]]
then
save_category=$line
else
echo $count "," $save_category "," $line
let count=$count+1
|
As I said before, that is not POSIX (actually, I think I said it
was not portable -- same thing).
count=$(( $count + 1 ))
| Quote: | fi
done < xxx.txt > new_one.txt
|
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence |
|
| Back to top |
|
|
jctown@nb.sympatico.ca
*nix forums beginner
Joined: 19 Jul 2006
Posts: 6
|
| Posted: Wed Jul 19, 2006 11:48 pm Post subject:
Re: joining lines in ksh
|
|
|
Chris F.A. Johnson wrote:
| Quote: | On 2006-07-19, jctown@nb.sympatico.ca wrote:
Ed Morton wrote:
Jon LaBadie wrote:
jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
My pseudo-code is not 100% accurate, but I think it's close enough.
As a ksh solution was requested:
while read rec
do
case "$rec"
in
category*) CAT="$rec" ;;
*) echo "$CAT,$rec" ;;
esac
done
Not a huge deal, but the normal convention for naming shell variables is
that all-upper-case is only used for exported variables. In the above
case, you run into the slight problem that if you make your variable
lower case then it becomes "cat" which clashes with the command name and
so could create some confusion to someone reading the script in future
and cause bizarre behavior if you use your variable incorrectly in a
context where "cat" gets interpretted as the command name. For example
if you mistakenly wrote this:
cat ="$rec"
instead of this:
cat="$rec"
then you'd get an error message about "cat" not being able to open the
file "=category1,member1-1".
To get round that, another convention that some people (apparently only
me in this NG so I often don't follow my own advice when posting here!)
use is to start all non-exported variables with an underscore as well as
make them lower (actually mixed) case. You also don't NEED the quotes
around "$rec" in the case statement. So, the above could be written as:
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed.
believe it or not, the solution was rather simple.
# count is a unique number so that I can use it as an index on a
database table
count=1
for line in $( cat text_file | awk '{print $1}' )
That is a useless use of cat, and an unnecessary use of awk. The
only point to using awk is to mitigate this incorrect method of
reading a file. Use:
while read line junk ## $junk will contain everything after the first word
do
echo $line | grep "a string"
There is no need for grep. Use case:
case $line in
*"a string"*) true ;;
*) false ;;
esac
if [[ $? -eq 0 ]]
Use [ $? -eq 0 ], as [[...]] is not portable.
then
# i need to save my category
save_category=$line
else
# build a line to put out
echo $count "," $save_category "," $line >> output_file
Your script will be faster if you put the redirection outside the
loop.
let count=$count+1
Use the portable form:
count=$(( $count + 1 ))
fi
done
done < text_file > output_file
It took a while to run, but I was able to create my file, 118000+ lines
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
|
Hi,
Here's my version of your script... my posix complained about [ $? -eq
0 ] so I put it back to [[ ]]
count=1
while read line junk ## $junk will contain everything after the first
word
do
case $line in
*".jar"*) true ;;
*) false ;;
esac
if [[ $? -eq 0 ]]
then
save_category=$line
else
echo $count "," $save_category "," $line
let count=$count+1
fi
done < xxx.txt > new_one.txt |
|
| Back to top |
|
|
jctown@nb.sympatico.ca
*nix forums beginner
Joined: 19 Jul 2006
Posts: 6
|
| Posted: Wed Jul 19, 2006 11:29 pm Post subject:
Re: joining lines in ksh
|
|
|
Chris F.A. Johnson wrote:
| Quote: | On 2006-07-19, jctown@nb.sympatico.ca wrote:
Ed Morton wrote:
Jon LaBadie wrote:
jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
My pseudo-code is not 100% accurate, but I think it's close enough.
As a ksh solution was requested:
while read rec
do
case "$rec"
in
category*) CAT="$rec" ;;
*) echo "$CAT,$rec" ;;
esac
done
Not a huge deal, but the normal convention for naming shell variables is
that all-upper-case is only used for exported variables. In the above
case, you run into the slight problem that if you make your variable
lower case then it becomes "cat" which clashes with the command name and
so could create some confusion to someone reading the script in future
and cause bizarre behavior if you use your variable incorrectly in a
context where "cat" gets interpretted as the command name. For example
if you mistakenly wrote this:
cat ="$rec"
instead of this:
cat="$rec"
then you'd get an error message about "cat" not being able to open the
file "=category1,member1-1".
To get round that, another convention that some people (apparently only
me in this NG so I often don't follow my own advice when posting here!)
use is to start all non-exported variables with an underscore as well as
make them lower (actually mixed) case. You also don't NEED the quotes
around "$rec" in the case statement. So, the above could be written as:
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed.
believe it or not, the solution was rather simple.
# count is a unique number so that I can use it as an index on a
database table
count=1
for line in $( cat text_file | awk '{print $1}' )
That is a useless use of cat, and an unnecessary use of awk. The
only point to using awk is to mitigate this incorrect method of
reading a file. Use:
while read line junk ## $junk will contain everything after the first word
do
echo $line | grep "a string"
There is no need for grep. Use case:
case $line in
*"a string"*) true ;;
*) false ;;
esac
if [[ $? -eq 0 ]]
Use [ $? -eq 0 ], as [[...]] is not portable.
then
# i need to save my category
save_category=$line
else
# build a line to put out
echo $count "," $save_category "," $line >> output_file
Your script will be faster if you put the redirection outside the
loop.
let count=$count+1
Use the portable form:
count=$(( $count + 1 ))
fi
done
done < text_file > output_file
It took a while to run, but I was able to create my file, 118000+ lines
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
|
Thank you so much for the advice...my script isn't elegant but it
works... it was a one-shot deal and not for public consumption...
However I will take what you have shown and give it a try as well.
Cheers |
|
| Back to top |
|
|
Chris F.A. Johnson
*nix forums Guru
Joined: 20 Feb 2005
Posts: 2268
|
| Posted: Wed Jul 19, 2006 8:45 pm Post subject:
Re: joining lines in ksh
|
|
|
On 2006-07-19, jctown@nb.sympatico.ca wrote:
| Quote: |
Ed Morton wrote:
Jon LaBadie wrote:
jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
My pseudo-code is not 100% accurate, but I think it's close enough.
As a ksh solution was requested:
while read rec
do
case "$rec"
in
category*) CAT="$rec" ;;
*) echo "$CAT,$rec" ;;
esac
done
Not a huge deal, but the normal convention for naming shell variables is
that all-upper-case is only used for exported variables. In the above
case, you run into the slight problem that if you make your variable
lower case then it becomes "cat" which clashes with the command name and
so could create some confusion to someone reading the script in future
and cause bizarre behavior if you use your variable incorrectly in a
context where "cat" gets interpretted as the command name. For example
if you mistakenly wrote this:
cat ="$rec"
instead of this:
cat="$rec"
then you'd get an error message about "cat" not being able to open the
file "=category1,member1-1".
To get round that, another convention that some people (apparently only
me in this NG so I often don't follow my own advice when posting here!)
use is to start all non-exported variables with an underscore as well as
make them lower (actually mixed) case. You also don't NEED the quotes
around "$rec" in the case statement. So, the above could be written as:
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed.
believe it or not, the solution was rather simple.
# count is a unique number so that I can use it as an index on a
database table
count=1
for line in $( cat text_file | awk '{print $1}' )
|
That is a useless use of cat, and an unnecessary use of awk. The
only point to using awk is to mitigate this incorrect method of
reading a file. Use:
while read line junk ## $junk will contain everything after the first word
| Quote: | do
echo $line | grep "a string"
|
There is no need for grep. Use case:
case $line in
*"a string"*) true ;;
*) false ;;
esac
Use [ $? -eq 0 ], as [[...]] is not portable.
| Quote: | then
# i need to save my category
save_category=$line
else
# build a line to put out
echo $count "," $save_category "," $line >> output_file
|
Your script will be faster if you put the redirection outside the
loop.
| Quote: | let count=$count+1
|
Use the portable form:
count=$(( $count + 1 ))
done < text_file > output_file
| Quote: | It took a while to run, but I was able to create my file, 118000+ lines
|
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence |
|
| Back to top |
|
|
jctown@nb.sympatico.ca
*nix forums beginner
Joined: 19 Jul 2006
Posts: 6
|
| Posted: Wed Jul 19, 2006 8:08 pm Post subject:
Re: joining lines in ksh
|
|
|
Jon LaBadie wrote:
| Quote: | jctown@nb.sympatico.ca wrote:
Ed Morton wrote:
Jon LaBadie wrote:
jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
...
What I need is a program that will read through the file and create a
new file based on
...
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
...
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed.
believe it or not, the solution was rather simple.
# count is a unique number so that I can use it as an index on a
database table
count=1
for line in $( cat text_file | awk '{print $1}' )
do
echo $line | grep "a string"
if [[ $? -eq 0 ]]
then
# i need to save my category
save_category=$line
else
# build a line to put out
echo $count "," $save_category "," $line >> output_file
let count=$count+1
fi
done
It took a while to run, but I was able to create my file, 118000+ lines
Glad you found a solution.
Yes it would take a while to run, you did 118000 creations of a pipe,
118000 start-up and runs of grep, 118000 openings of output_file.
Your original requirement did not show any need to extract column
one from the input, in fact the sample data only had one column.
Nor did you mention the need to add a counter.
Should you have to repeat the run, I think you would find considerable
speed-up by doing those actions once, or in the shell rather than a
separate program.
Using Ed's code as a base:
count=0
while read rec
do
case $rec in
*a_string*) _cat="$rec" ;;
*) let count=$count+1
echo "$count,$_cat,$rec"
;;
esac
done < text_file > output_file
And if you need the first column extraction, a small change:
awk '{print $1}' text_file |
while read rec
...
done > output_file
|
There is only one column in the input file, I just want to make sure I
didn't get any other stuff on the line that might confuse the issue. |
|
| Back to top |
|
|
Jon LaBadie
*nix forums beginner
Joined: 06 May 2006
Posts: 46
|
| Posted: Wed Jul 19, 2006 4:56 pm Post subject:
Re: joining lines in ksh
|
|
|
jctown@nb.sympatico.ca wrote:
| Quote: | Ed Morton wrote:
Jon LaBadie wrote:
jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
....
What I need is a program that will read through the file and create a
new file based on
....
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
....
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed.
believe it or not, the solution was rather simple.
# count is a unique number so that I can use it as an index on a
database table
count=1
for line in $( cat text_file | awk '{print $1}' )
do
echo $line | grep "a string"
if [[ $? -eq 0 ]]
then
# i need to save my category
save_category=$line
else
# build a line to put out
echo $count "," $save_category "," $line >> output_file
let count=$count+1
fi
done
It took a while to run, but I was able to create my file, 118000+ lines
|
Glad you found a solution.
Yes it would take a while to run, you did 118000 creations of a pipe,
118000 start-up and runs of grep, 118000 openings of output_file.
Your original requirement did not show any need to extract column
one from the input, in fact the sample data only had one column.
Nor did you mention the need to add a counter.
Should you have to repeat the run, I think you would find considerable
speed-up by doing those actions once, or in the shell rather than a
separate program.
Using Ed's code as a base:
count=0
while read rec
do
case $rec in
*a_string*) _cat="$rec" ;;
*) let count=$count+1
echo "$count,$_cat,$rec"
;;
esac
done < text_file > output_file
And if you need the first column extraction, a small change:
awk '{print $1}' text_file |
while read rec
...
done > output_file |
|
| Back to top |
|
|
jctown@nb.sympatico.ca
*nix forums beginner
Joined: 19 Jul 2006
Posts: 6
|
| Posted: Wed Jul 19, 2006 3:49 pm Post subject:
Re: joining lines in ksh
|
|
|
Ed Morton wrote:
| Quote: | Jon LaBadie wrote:
jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
My pseudo-code is not 100% accurate, but I think it's close enough.
As a ksh solution was requested:
while read rec
do
case "$rec"
in
category*) CAT="$rec" ;;
*) echo "$CAT,$rec" ;;
esac
done
Not a huge deal, but the normal convention for naming shell variables is
that all-upper-case is only used for exported variables. In the above
case, you run into the slight problem that if you make your variable
lower case then it becomes "cat" which clashes with the command name and
so could create some confusion to someone reading the script in future
and cause bizarre behavior if you use your variable incorrectly in a
context where "cat" gets interpretted as the command name. For example
if you mistakenly wrote this:
cat ="$rec"
instead of this:
cat="$rec"
then you'd get an error message about "cat" not being able to open the
file "=category1,member1-1".
To get round that, another convention that some people (apparently only
me in this NG so I often don't follow my own advice when posting here!)
use is to start all non-exported variables with an underscore as well as
make them lower (actually mixed) case. You also don't NEED the quotes
around "$rec" in the case statement. So, the above could be written as:
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed.
|
believe it or not, the solution was rather simple.
# count is a unique number so that I can use it as an index on a
database table
count=1
for line in $( cat text_file | awk '{print $1}' )
do
echo $line | grep "a string"
if [[ $? -eq 0 ]]
then
# i need to save my category
save_category=$line
else
# build a line to put out
echo $count "," $save_category "," $line >> output_file
let count=$count+1
fi
done
It took a while to run, but I was able to create my file, 118000+ lines |
|
| Back to top |
|
|
Ed Morton
*nix forums Guru
Joined: 20 Feb 2005
Posts: 1073
|
| Posted: Wed Jul 19, 2006 1:47 pm Post subject:
Re: joining lines in ksh
|
|
|
Jon LaBadie wrote:
| Quote: | jctown@nb.sympatico.ca wrote:
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
My pseudo-code is not 100% accurate, but I think it's close enough.
As a ksh solution was requested:
while read rec
do
case "$rec"
in
category*) CAT="$rec" ;;
*) echo "$CAT,$rec" ;;
esac
done
|
Not a huge deal, but the normal convention for naming shell variables is
that all-upper-case is only used for exported variables. In the above
case, you run into the slight problem that if you make your variable
lower case then it becomes "cat" which clashes with the command name and
so could create some confusion to someone reading the script in future
and cause bizarre behavior if you use your variable incorrectly in a
context where "cat" gets interpretted as the command name. For example
if you mistakenly wrote this:
cat ="$rec"
instead of this:
cat="$rec"
then you'd get an error message about "cat" not being able to open the
file "=category1,member1-1".
To get round that, another convention that some people (apparently only
me in this NG so I often don't follow my own advice when posting here!)
use is to start all non-exported variables with an underscore as well as
make them lower (actually mixed) case. You also don't NEED the quotes
around "$rec" in the case statement. So, the above could be written as:
while read rec
do
case $rec in
category*) _cat="$rec" ;;
*) echo "$_cat,$rec" ;;
esac
done
Regards,
Ed. |
|
| Back to top |
|
|
Jon LaBadie
*nix forums beginner
Joined: 06 May 2006
Posts: 46
|
| Posted: Wed Jul 19, 2006 5:16 am Post subject:
Re: joining lines in ksh
|
|
|
jctown@nb.sympatico.ca wrote:
| Quote: | I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
My pseudo-code is not 100% accurate, but I think it's close enough.
|
As a ksh solution was requested:
while read rec
do
case "$rec"
in
category*) CAT="$rec" ;;
*) echo "$CAT,$rec" ;;
esac
done |
|
| Back to top |
|
|
John W. Krahn
*nix forums Guru
Joined: 27 Feb 2005
Posts: 602
|
| Posted: Wed Jul 19, 2006 4:30 am Post subject:
Re: joining lines in ksh
|
|
|
jctown@nb.sympatico.ca wrote:
| Quote: | I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
[snip]
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
|
$ echo "category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2" | perl -lne'/^category/?($cat=$_):print"$cat,$_"'
category1,member1-1
category1,member1-2
category1,member1-3
category1,member1-4
category1,member1-5
category1,member1-6
category2,member2-1
category2,member2-2
category2,member2-3
category2,member2-4
category2,member2-5
category2,member2-6
category3,member3-1
category3,member3-2
John
--
use Perl;
program
fulfillment |
|
| Back to top |
|
|
Chris F.A. Johnson
*nix forums Guru
Joined: 20 Feb 2005
Posts: 2268
|
| Posted: Wed Jul 19, 2006 1:55 am Post subject:
Re: joining lines in ksh
|
|
|
On 2006-07-19, jctown@nb.sympatico.ca wrote:
| Quote: | I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
.
.
category2,member2-1
category2,member-2-
.
.
|
awk '/^category/ {category = $0; next}
{ print category "," $0 }' FILENAME > NEWFILE
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence |
|
| Back to top |
|
|
jctown@nb.sympatico.ca
*nix forums beginner
Joined: 19 Jul 2006
Posts: 6
|
| Posted: Wed Jul 19, 2006 1:28 am Post subject:
joining lines in ksh
|
|
|
I have a file that looks like this
category1
member1-1
member1-2
member1-3
member1-4
member1-5
member1-6
category2
member2-1
member2-2
member2-3
member2-4
member2-5
member2-6
category3
member3-1
member3-2
I have 432 categories
What I need is a program that will read through the file and create a
new file based on
the following pseudo code
while not eof input
read a line
if it's a category
save the category
endif
read a line
if it's not a category
create a new line with saved category,member
end
Which would result in something line this
category1,member1-1
category1,member1-2
..
..
category2,member2-1
category2,member-2-
..
..
My pseudo-code is not 100% accurate, but I think it's close enough. |
|
| Back to top |
|
|
Google
|
|
| Back to top |
|
|
|
|
The time now is Fri Nov 21, 2008 7:08 pm | All times are GMT
|
|
Mobile Phones | Credit Cards | Outsourcing | Loans | Credit Card
|
|
Copyright © 2004-2005 DeniX Solutions SRL
|
|
|
|
Other DeniX Solutions sites:
Unix/Linux blog |
electronics forum |
medicine forum |
science forum |
|
|
Privacy Policy
|
Powered by phpBB © 2001, 2005 phpBB Group
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Forum index
»
Programming
»
shell
