I'm looking for the best way to display all rows from two tables while joining first by one field (dwg) then where applicable a 2nd join on part.

Table1 data consists of schematics(dwg) along with a list of parts required to build the item depicted in the drawing. Table2 consists of data about the actual parts ordered to build the schematic. Some parts in table2 are a combination of parts in table1 (ex: foo and bar in table1 were ordered as foobar in table2).

I can display all rows in both tables with UNION ALL, but this doesn't join on both the dwg and part fields. I looked at FULL OUTER JOIN also, but I haven't figured out how to join first by dwg, then by part. Here is an example of the data.

table1                   table2   
dwg        part    qty       order   dwg      part    qty  
-----      -----   -----    -----    -----   -----   -----  
123        foo     1        ord1     123     foobar  1  
123        bar     1        ord1     123     bracket 2  
123        widget  2        ord2     123     screw   4  
123        bracket 4        ord2     123     nut     4  
456        foo     1        ord2     123     widget  2  
                            ord2     123     bracket 2  
                            ord3     456     foo     1  

Desired output:
The goal is to create a view that provides visibility to all parts in table1 and the associated orders in table2 (including those parts that appear in one but not the other table) so that I can see all the drawing parts in table1 and the associated records in table2 along with records in table2 where the part wasn't in table1.

part_request_order_report  
dwg      part      qty   order   part    qty  
-----    -----     ----- ------  -----   -----  
123      foo       1  
123      bar       1  
123      widget    2     ord2    widget  2  
123      bracket   4     ord1    bracket 2  
123      bracket   4     ord2    bracket 2  
123                      ord1    foobar  1  
123                      ord1    screw   4  
123                      ord1    nut     4  
456      foo       1     ord3    foo     1  

Is this possible? Or am I better off iterating through the data to build the report table? Thanks in advance.