Outline of One Possible Solution to Programming Assignment 2.
=============================================================
Note that other designs are possible, the critical thing is
to understand the tradeoffs. This outline only lists the
high level steps, it does not structure the code nicely into
objects/subroutines etc.
=============================================================

connlist : array of connections
rclist   : array of return codes
tlist    : array of threads

Read config file and sql file

---------------------------------------
Part 3: Naive Distributed SQL Processor
---------------------------------------

parse SQL

if SQL is create table statement 
then 
   createTable(cfgfile, sql, connlist, rclist, tlist) from assignment 1
elsif SQL is drop table statement
then
   dropTable(cfgfile, sql, connlist, rclist, tlist) from assignment 1
elsif SQL is select-from-where statement
then 
   processQuery(cfgfile, sql, connlist, rclist, tlist) 
endif

PROCEDURE processQuery(cfgfile, sql, connlist, rclist, tlist) 
-------------------------------------------------------------

query catalog db for node & connection info for given table name

foreach node n, connlist[n] = open connection.

foreach node n, tlist[n] = runThread (sql, rclist, n)
runThread:
   run SQL query
   while(resultset not empty)
      fetch results
      foreach column
         print results

Wait for each thread to finish

foreach node n check rclist[n] and output success or failure

close all node and catalog connections

perform other clean up

---------------------------------------
Part 4: Naive Distributed SQL Processor
---------------------------------------

Read config file 

query catalog db for node & connection info for given table name

update catalog db with partition information

connlist = open connections such that nodeid is the index into the array

if partmtd=notpartition

   foreach line in csv file
      foreach nodeid  associated with the table
         insert(connlist[nodeid], line, rclist[i]) 

elsif partmtd=hash

   foreach line in csv file
      partcolval = get partitioning column(s)
      nodeid = hashfunction(partcolval, partparam1)
      insert(connlist[nodeid], line, rclist[nodeid]) 

elsif partmtd=range
   partition ranges should be read into an in-memory data structure.   
   foreach line in csv file
      partcolval = get partitioning column(s)
      nodeid = findrange(partcolval, ranges2nodeid)
      insert(connlist[nodeid], line, rclist[nodeid]) 
else
   error

foreach node n check rclist[n] and output success or failure, number of rows inserted, number of rows rejected etc. 

close all node and catalog connections

perform other clean up