UNIX for Dummies Questions & Answers

Dear unix gurus,

I have a data file with header information about a subject and also 3 columns of n rows of data on various items he owns. The data file looks something like this:

adam peter
blah blah blah
blah blah blah

car

01 30 200
02 31 400
03 57 121
.. .. ..
.. .. ..
n y z

bicycle

01 21 888
02 25 213
03 35 414
.. .. ...
n y z

van

... ... ...
... ... ...
n x y

What I would like to do is write a script to loop over all the items i.e. car, bicycle and van, create a new file named after the item and pipe in all the 3 columns X n rows of data associated with each item into the file. So I should end up with 3 files called car, bicycle, van, each with its associated data only. Every item has exactly 3 columns and n rows of data.
I'm struggling...can someone help? thanks a lot.

Code:

awk 'NF == 1 {n=$1} NF > 1 {print $0 >> n}' infile

thanks. but your suggestion did not help. The first thing is that I get the error message "( FILENAME=xxx.dat FNR=1) fatal: expression for >> has null string value"
The second this is that it seems that you are trying to pipe the contents of $0 into n. In which case we would end up with only a single file. What I want is 3 separate files e.g

file 1: car

01 30 200
02 31 400
03 57 121
.. .. ..
.. .. ..
n y z

file 2 :car

01 30 200
02 31 400
03 57 121
.. .. ..
.. .. ..
n y z

and so on. thanks.

No, sometimes I even know what I do Usually I test a solution out before I post it. I am usually trying not to guess by the description the posters give.

The thing is that you might use another version of awk than I do. I use gawk.
A quote I always have to search which might help you:
To explain what it does:
If the line has only 1 field it will assing field 1 to n. If the line has more than 1 field, it will write the line to the file named n. It does that for every line.

Here is what I get with the solution I posted - I altered your example a bit but it stays the same like what you posted as input:

Code:

$> cat infile2
car

 1 2 3
 4 5 6
 7 8 9

bicycle

 10 11 12
 13 14 15
 16 17 18

van

 21 22 23
 24 25 26
 27 28 29
$> awk 'NF == 1 {n=$1} NF > 1 {print $0 >> n}' infile2
$> ll car bicycle van
-rw-r--r-- 1 root root 30 2009-03-27 09:30 bicycle
-rw-r--r-- 1 root root 21 2009-03-27 09:30 car
-rw-r--r-- 1 root root 30 2009-03-27 09:30 van
$> cat bicycle
 10 11 12
 13 14 15
 16 17 18
$> cat car
 1 2 3
 4 5 6
 7 8 9
$> cat van
 21 22 23
 24 25 26
 27 28 29

The output is 3 different files each containing the rows that are contained under the "filename" in the single original file.

Also please use code tags next time posting code, logs or data etc.

yes, that is exactly what I want. I think the problem is that we are using different platforms. the quote use nawk doesn't work in my shell.

My shell type is bash-3.2. my GNU awk version is 3.1.6. Same applies for gawk. Would you know what the equivalent command of your script is for bash shell?

The solution of zaxxon should work with gawk. Post a part of your input file within code tags, select the text with your mouse and click on the # above the edit box.

Regards

I'm not sure what you mean Franklin52, I'm new in the forum. How will wrapping my input file within code tags help me?

Code:

 
peter adams
40 the for five
car
01 1.1 1.96
02 31 55.2
03 29 23.4
04 27 24.7
05 25 54
06 23 55
 
bicycle
 
01 42 100
02 32.1 44.66
03 33 44.1
04 45.8 54.2
05 66 53.1
06 48 22

what do I do next?