Something like the below makes sense to me as your data structure. Keying the outside hash to the common key for your tables allows you to add data to the anonymous hashes inside as you find new records for an employee, rather than having to put an employee record together completely before storing it in the data structure. Then, just walk over the outside hash and output the key and the data from inside the anonymous hash for each employee in turn.

@employee = (
    '227' => {
        'uid' => 'simpsonh',
        'sn' => 'Simpson',
        'givenname' => 'Homer',
        'manager' => 'burnsm',
        'telephoneNumber' => '2183445555',
        'departmentNumber' => 38,
        'title' => 'Nuclear Plant Operator',
        'roomNumber' => 7,
        'buildingName' => 'Nuclear Resources'
    },
    222 => {
       # rest of entry info here
    }
);
[download]

What do you want to sort on? That would indicate how to store the data when you load it.

In any event, use a hash keyed on whatever field you decide, and use the Perl sorted keys idiom when you loop through the structure.

OK, apparently I am going to have to re-write most of the script now because of your suggestion. This is a good thing though because I was able to get the MySQL data into a nicely formatted, comma-delimited csv file.

This is the select statement I used, if anyone is wondering:

SELECT Employee.username, Employee.first_name, Employee.last_name, Dep
+artment.name, LocationBuilding.name, PhoneNumbers.phone_id, PhoneNumb
+ers.fax_flag, PhoneNumbers.exchange, P.position, OrgChart.supervisor_
+id FROM P INNER JOIN OrgChart ON(P.employee_id=OrgChart.employee_id) 
+INNER JOIN Employee ON(OrgChart.employee_id=Employee.employee_id) INN
+ER JOIN PhoneNumbers ON(Employee.employee_id=PhoneNumbers.employee_id
+) INNER JOIN Department ON(OrgChart.department_id=Department.departme
+nt_id) INNER JOIN LocationBuilding ON(Department.building_id=Location
+Building.building_id) INNER JOIN Location ON(LocationBuilding.buildin
+g_id=Location.building_id) INNER JOIN LocationRooms ON(Location.room_
+id=LocationRooms.room_id);" | sed 's/\t/","/g;s/^/"/;s/$/"/;s/\n//g' 
+> Employee.csv
[download]

Thanks!

employee_info
managers
buildings
locations
positions
rooms
departments

And you know, it appears, that data grouped on successive lines relate to each other and that groupings are separated by blank lines, and that data tags and their value are spearated by a colon. So, go through the file line by line, read data into a buffer by reading forward until you hit a blank line, then store that into temporary hashes as if in sql tables. Once it's all compiled, then output your total file.

pseudo-code:

open FILE, $infile;
@lines = <FILE>;
close FILE;

for $i (0 .. $#lines) {
    if ($lines($i) ne "") {
        ($tag, $value) = split(/:/, $lines($i));
        $temp{$tag} = $value;
    }
    else {
        #reached end of a data grouping, now process it
        if exists $temp{'employee_id'} {
            foreach $field (keys %temp) {
                $employee_info{$field} = $temp{$field};
            }
        }
        elseif (exists $temp{'position_id') && exists $temp{'title'}) 
+{
            $positions{$temp{'position_id')} = $temp{'title'};
        }
        elseif {
        ....etc...
        }
    ...etc...
    }
    ...clear %temp...
}

foreach $employee (keys %employee_info) {
    output info, drawing labels from %positions, %departments, etc...
}
[download]