Removing specific lines from a file

[complexkid]

Hi,
I have a huge file having ~500,000 records. I want to remove certain records satisfying some criteria (that field 3 and 4 both have value 0). I wrote following script, but I am getting errors:

#!/bin/sh
lno=0
while read line
do
lno=`expr $lno + 1`
echo "line read" $lno
annot =`echo $line | cut -f 3 -d " "`
gen = `echo $line | cut -f 4 -d " "`

if [ $annot -eq 0 ]
then
if [ $gen -eq 0 ]
then
echo "delete it"
fi
else
echo $line >> balancedset
fi
done < ./newTrainingset12.dat
-----------------------------------------------#
Errors:
removeoutliers.sh: line 8: annot: command not found
removeoutliers.sh: line 9: gen: command not found
removeoutliers.sh: line 11: [: -eq: unary operator expected


-----
I tried doing the command substitution on the prompt, and it works there.
could anyone tell me what's wrong here?
Thanks

[drl]

Hi.

Do not use whitespace around "=" in assignment statements.

That's a common typo for folks who go back and forth among many languages.

On the other hand, you must supply whitespace around "[" and "]", and the relational operators, like "=", in the if statements.

Hang in there, your fingers will soon learn what's right where ... cheers, drl

[scm]

And you can short-circuit some of your tests:

if [ $annot -eq 0 -a $gen -eq 0 ] 
then
...

Don't forget to truncate your output file before you start, in case you rerun your script:

>balancedset

will do it, before you start your "while read line" loop.

[ghostdog74]

for serious parsing and efficiency only

awk '$3 != 0 && $4 != 0' file  > newfile