[SOLVED] awk delimiters

Philip Kiser · 08-18-2010, 10:25 AM

Hi,

Let me apologize in advance for the simple question.

I have the following text.

PHP Code:



1234567890123456789012 
 
1  N   AAA  45  65 231 
2  FGH AAA  40  541452 
3  SGFEAAA  55  45 735 
4  DSG BBB  62  54 552 
5  GHH CCC  45  87 664

The first line of numbers is to indicate the position of the text below. The text has six columns of information: A number that occupies positions 1-3,
a name that occupies positions 4-7, an alternative name that occupies positions
8-10 and three numbers that occupy positions 13-14, 17-18 and 19-22, respectively.

I would like to use awk to select records that have AAA as their
alternative name (records 1-3) and then calculate the sum of the
numbers that occupy positions 19-22 in these records (231+1452+735)
As you can see, some of the records are not separated by white space so
I can't use the default awk delimiter.

My question is how to set up the awk field delimiters to accomplish
this task.

Many thanks for your help.

Philip

grail · 08-18-2010, 10:48 AM

Well awk is generally devised around having a constant delimiter, such as a space. Seeing as there is no real pattern and you wish to use
awk I would suggest setting it to empty which will then make every character equal to a field:

Code:

awk -F "" '<yourstuff>' file

Please note the space between the F and first inverted commas "

Philip Kiser · 08-18-2010, 10:57 AM

Grail,

Thanks a lot. I didn't know that was a possibility.

Best wishes,
Philip

grail · 08-18-2010, 11:14 AM

All good

I find this site very useful