awk help wanted

db_Romain · #1 (**permalink**) 03-13-04, 09:14

Hi, I am new to bash programming.

I have different tab-separated text files with several lines each having colums that look like that :

ID Designation Price Stock
065 Soap 2.00 30
780 Parachute 56.00 2

Each file can have different field names and different number of fields.

I would like to retieve the field names and create variables ,one for each field ,with a name corresponding to a field name and whose value would be the field number.(e.g ID=1,Stock=4...)
I would then like to use these variables in the rest of my script
I came close using an awk but couldn't export anything back to the script

Any help would be greatly appreciated

aigles · #2 (**permalink**) 03-13-04, 10:52

You can do something like this :

Code:

#
# Function : FileFields input_file ...
# Output   : Fields assignments for each field of each record
#
FileFields() {
   awk '
        # First record of input_file contains field names
        # which are memorized in the col[] array.
        FNR==1 {
           col_count=NF;
           for (f=1; f<=NF; f++)
              col[f] = $f;
           next;
        }
        # Data record
        # Assignment for each file is generated in the form :
        # field_name='field_value';
        {
           for (f=1; f<=col_count; f++)
             printf("%s=\047%s\047; ", col[f], $f);
             print "";
        }
       ' $*
}

#
# Proceed input files
# The input files are read with the FileFields function
#
FileFields cols.txt | \
while read fields
do
   eval $fields

   # Proceed record
   # Displays assignments of variables as an example
   echo $fields

done

The result of the execution of this script with your input datas is :

ID='065'; Designation='Soap'; Price='2.00'; Stock='30';
ID='780'; Designation='Parachute'; Price='56.00'; Stock='2';

db_Romain · #3 (**permalink**) 03-13-04, 11:30

Thanks for your help, but that's not exactly what i meant

what i would like is to declare variables in my bash script that would have the name of a field and the position in the inventory file as value

for example ,with this inventory file(tab-separated):
ID Designation Price Stock
065 Italian Wine 2.00 30
780 Parachute 56.00 2

i would like to be able to use in the rest of my script (and outside the awk) the variables ID,DESIGNATION,PRICE,STOCK which values are:
ID=1
DESIGNATION=2
PRICE=3
STOCK=4
instead of having to declare them

for this other inventory file:
Name Surname Age
Keeler John 41
King Peter 32

i would like to be able to use in the rest of my script the variables
NAME=1
SURNAME=2
AGE=3

Sorry for the the misunderstanding.Your help would be greatly appreciated because i can't find the solution and it's driving me crazy!

aigles · #4 (**permalink**) 03-13-04, 11:57

Try this new script :

Code:

#
# Function : GetFields - Get field names and positions
# Args(s)  : $1 = Input file name (Record 1 = Field names)
#            $2 = Prefix for variable names (Def="")
# Env.     : A variable is created for each field of the file
#            in the form <Prefix><FIELDNAME> and is assigned
#            the field position.
#
GetFields() {
   local fields=$( awk '
        NR==1 {
           for (field=1; field<=NF; field++)
              printf("%s%s=%d;", PREFIX, toupper($field), field);
           print "";
           exit;
        }
       ' PREFIX="$2" $1 )
   eval $fields
}

#
# M A I N . . .
#
GetFields input_field FIELD_
set | grep '^FIELD_'

The result for your inventory file is :

FIELD_DESIGNATION=2
FIELD_ID=1
FIELD_PRICE=3
FIELD_STOCK=4

In my example, th variable are prefixed by FIELD_
This prefix is optional.

db_Romain · #5 (**permalink**) 03-13-04, 12:23

Thanks!!!
I don't think i can be of any assistance to you in bash scripting
but i really owe you one

If i encounter another problem that's definitely here that i'll post it

thanks again!

Internet Services

Web Development

Digital Marketing

Consumer Tech