asked on

Search a specific field in a table

Hi,

I'm trying to use grep to searh for a string in a table. However I need to limit the search to a specific field as I dont want any results from the first field if it contains the string. Then I need to print a few of the fields from the lines that contain my string in the field I need to search on. Here's an example of my data:

Toteve Sam Denver
Denver John Aspen
Ellis Joe Witchita
Toms Mary Denver

I want to search column 3 for Denver, but not pick up the John Denver line as I'm only searching column 3. It's easy in awk, but I cant use awk... Do I use Grep....I dont see how I can use Grep.

Thanks,

Sam

tfewster

grep "Denver$" yourfile

Tintin

You "can't use awk". Warning! Warning! Homework question.

elwayisgod

ASKER

Why does this not work?

#!/usr/bin/ksh

# Search by Name

echo "What Last Name do you want to search for: \n"
read lname

nawk -F"," '$1 == /$lname/ { print }' flight.data

tfewster

awk doesn't accept global parameter substitution in that way; You have pass parameters from the awk command line, e.g. nawk -F"," '$1 == NAME { print }' -v NAME=$lname flight.data

It occurred to me that you might be using grep because you didn't know how to do case-insensitive comparisons in awk; The answer to that is, use IGNORECASE, e.g.
lname=denVer
nawk -F"," 'BEGIN {IGNORECASE=1} $1 == NAME { print }' -v NAME=$lname flight.data

(P.S. Tintin - I agree it looks odd, but checking the profile reassured me ;-)

elwayisgod

ASKER

Not sure about the -v option. I think I can do the NAME=$lname before the print statement. However, can't figure out the -v. Here is my current script. If i leave the -v out, it does not print the lines where 'Toteve' is present. It is just blank output.

#!/usr/bin/ksh

# Search by Name

echo "What Last Name do you want to search for: \n"
read lname
echo $lname
NAME=$lname

nawk -F"," '$1 == NAME { print }' -v flight.data

tfewster

Did you try it the way I suggested?
nawk -F"," '$1 == MYAWKVAR { print }' -v MYAWKVAR=$lname flight.data

If you don't do the assignment on the awk command line, MYAWKVAR is undefined within awk, even it it's a global variable. You're using Solaris, aren't you? Check the -v option in the man page for nawk.

elwayisgod

ASKER

Here is my script and the results of output when I do it the way you indicated. It thinks -v is a file, thus errors out...

SCRIPT:
#!/usr/bin/ksh

# Search by Name

echo "What Last Name do you want to search for: \n"
read lname
echo $lname
NAME=$lname
echo $NAME

nawk -F"," '$ 1== NAME { print }' -v NAME=$lname flight.data

OUTPUT:

broncos:/Users/stoteve/regis/project 73 % ./name.ksh
What Last Name do you want to search for:

Toteve
Toteve
Toteve
nawk: can't open file -v
source line number 1
broncos:/Users/stoteve/regis/project 74 %

ASKER CERTIFIED SOLUTION

tfewster

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

elwayisgod

ASKER

Shouldnt this sort correctly. When I run and choose to sort by option 3, I get:

SMITH,BRAD,STL,DEN,147,DELAYED,05/11/2004
SMITH,ROD,DEN,LAX,214,ONTIME,08/21/2004

Why is DEN line not above the STL line?

#!/usr/bin/ksh

# Search by Name

cat flight.data | tr [a-z] [A-Z] > flight.data #Change data file to UPPERCASE

echo "What Last Name do you want to search for: \n"
read lname

echo "$lname" | tr [a-z] [A-Z] > lnametr
cat lnametr
NAME=`cat lnametr`

echo "What field would you like to sort your results by: \n\n
      1) Last Name \n
      2) First Name \n
      3) Departure City \n
      4) Arrival City \n
      5) Flight Number \n
      6) Departure Date \n"

read srtfld

#cat newflight.data

case "$srtfld" in

      1)cat newflight.data | sort +0 newflight.data
            ;;
      2)cat newflight.data | sort +1 newflight.data
            ;;
      3)cat newflight.data | sort +2 newflight.data
            ;;
      4)cat newflight.data | sort +3 newflight.data
            ;;
      5)cat newflight.data | sort +4 newflight.data
            ;;
      6)cat newflight.data | sort +6 newflight.data
            ;;
      *) echo "Invalid Option. Sorting by default 'Last Name' \n"
             cat newflight.data | sort +0 newflight.data
            ;;

esac

# Line 47
nawk -F"," -v NAME=`cat lnametr` '$ 1== NAME { print }' flight.data > newflight.data
#cat newflight.data | sort +$srtfld newflight.data

elwayisgod

ASKER

Nevermind, used the -t, option as it was comma delimited..

sam

tfewster

Hmmm - That has nothing to do with the awk/grep questions and it's a bit unreasonable. Please see https://www.experts-exchange.com/Operating_Systems/Unix/help.jsp#hi107

However, you need to specify the delimiter character to `sort` using the -t option; As it is, it sees each line as a single field, and this cannot sort on "field 3".

Furthermore, `cat newflight.data | sort +6 newflight.data` is redundant; Just do `sort -t, +6 newflight.data`

Finally, newflight.data doesn't get created until the end of the script "#Line 47", so the case statement must be working on old data.

Tintin

grep is still suitable (and easier)

#!/usr/bin/ksh
echo "What Last Name do you want to search for: \c"
read lname

grep "^$lname" flight.data

For your sorting script, it would be better written as:

#!/usr/bin/ksh

echo "What Last Name do you want to search for: \c"
read lname

cat <<EOF
What field would you like to sort your results by:

1) Last Name
2) First Name
3) Departure City
4) Arrival City
5) Flight Number
6) Departure Date

EOF

echo "Enter option: \c"
read srtfld

case "$srtfld" in
1) sort -t, +0 newflight.data ;;
2) sort -t, +1 newflight.data ;;
3) sort -t, +2 newflight.data ;;
4) sort -t, +3 newflight.data ;;
5) sort -t, +4 newflight.data ;;
6) sort -t, +6 newflight.data ;;
*) echo "Invalid Option. Sorting by default 'Last Name'"
sort -t +0 flight.data
;;
esac

grep -i "^$lname" newflight.data