filtering input from read command

Shell Programming and Scripting
1

I need help understanding a script I'm modifying which someone else has written. Basically I�m looping through a buffer that holds records fetched from a database query. I need a way to separate the primary key values from other attributes in the result. Heres the code:

BUFF=buffer_file
[query results] > $BUFF
{
   while read row
   do
       rowID= $(row#*|)   #??
       rowCol=$(row%*|)        #??
   done
} < $BUFF
2

I think the last > is probably supposed to be a <.

this table of string operations explains what those mean among many more, assuming you're using a bourne or ksh shell.

Assuming the number of columns is always the same, there's probably a better way to do this too.

3

I agree

4

Is the number of columns always the same, though? Hard to offer suggestions when I don't know what the input and output is supposed to be.

5

Sorry for not specifying. I'm using bourne and yes the collumns are always the same (only two). Output from the query would look something like this:
1 server001
2 server002
3 server003

---------- Post updated at 12:47 PM ---------- Previous update was at 12:45 PM ----------

I seem to have found a solution that has been working but may not be the best immplementation:

rowID=$(echo $row | awk '{print $1}')
rowCol=$(echo $row | awk '{print $2}')
6

Easy peasy. read does this all by itself.

while read rowID rowCol
do
        echo "rowID=$rowID rowCol=$rowCol"
done < filename

If the separator was | instead of whitespace you'd just do

while IFS="|" read rowID rowCol
do
        echo "rowID=$rowID rowCol=$rowCol"
done < filename

This builtin is hundreds of times faster than calling awk for individual lines.