I think you already get the best advice possible in terms of implementation, so i will concentrate on the intrinsics of the problem:

Basically you have to parse a file, do a group change and reduce your input set by a set of key values. All these procedures/algorithms are basic tasks which you will have to handle on a daily basis in one or the other form in your future job. Probably this is why your professor designed the task this way.

Here is what a group change is and how it is carried out:

You have a file/table/datastructure with basically two columns: a "key" column and a "value" column. You sort that file for the key, then read it line by line and every time the key value changes you have to carry out some action.

For instance: Every customer of a business, instead of paying, will have written a record in a book (=file): his name and the value of the good he takes. At the end it looks like:

Code:

customerA  50,-
customerA  25,-
customerB  70,-
customerA  15,-
customerC  60,-
customerB  25,-
...

The task is to sum up the total for every customer and print a list with the totals. This is how it is done:

First sort the file for the key value, which is the first field in this case (the customer name):

Code:

customerA  50,-
customerA  25,-
customerA  15,-
customerB  70,-
customerB  25,-
customerC  60,-
...

Now read that file line by line. Store the last value of your key into a buffer. Optionally you might have an default action for every line, which be carried out here - in this case sum up the value. Whenever the key changes the comparison of the new key to the last key in this buffer will fail. Then you have to do your action (in this case print the total) and start over with the new key stored in the buffer. (Don't forget that at files end you have to end-process your last key value.)

Here is it in pseudo-code:

Code:

LAST=""
while (lines to process)
     KEY := KEY value of the line
     VAL := VALUE value of the line
     line_process( KEY, VALUE )
     if ( KEY differs from LAST )
          group_end_process( KEY [, ..... ] )
          LAST := KEY
     endif
     next line
end while
end_process( KEY [, ..... ] )

This is called a "single group change". There are also multiple group changes: suppose the customers would want a detailed statement, where the goods are sorted in groups and a subtotal should be given for every group of goods and a grand total for the customer (dual group change). It will need a second key field and i am sure you could write the pseudo code for this one already yourself.

I hope this helps in understanding.

(corollary: read carefully the man page of "awk", especially the part about how awk works and how "awk" scripts are formed. Does that ring a bell?)

bakunin