22S:172 Summer 2005
Midterm

This exam requires the use of two data files, which may be downloaded
from the course web page under "Datasets."  They are

    customers.dat
    invoices.dat

The "customers.dat" file contains information on some of the customers of 
Hank's Hardware.  The variables are:

   customer id number
   customer name
   customer type ( C = commercial; I = individual )
   customer contact date  

This is what the file looks like.

001 Johnson Floral   C   01/02/2002
002 Wang Jian        I   01/15/2002
003 TrueSale Jewelry C   06/25/2003
004 Salo Farms       D   09/18/2003
005 Jackson Michael  I   12/15/2003
006 Armstrong Lance  I   01/13/2004
007 Askov Lighting   C   12/14/2004

The "invoices.dat" file contains summary information on invoices processed by
Hank's Hardware.  The variables are:

   invoice number
   customer id number
   invoice date
   invoice total

This is what this file looks like.

0001 002  05/15/2004  35.00
0002 002  06/12/2004  25.50
0003 001  07/15/2004   5.00
0004 002  12/15/2004  72.50
0005 003  01/25/2005  10.00
0006 005  03/15/2005  52.75
0007 002  04/12/2005  25.00
0008 001  07/12/2000  16.00

Here are your tasks:

1.  Write a SAS data step that will read the customers file. Hint: The 
    dates in the file are in mmddyy10. format.

2.  Write a SAS procedure to produce the following table of counts of the
    number of each unique value of customer type.

                           custtype    Frequency
                           ---------------------
                           C                  3 
                           D                  1 
                           I                  3 

3.  Write a SAS data step that will read the invoices file.

4.  Create a new dataset by merging the customers file with the invoices
    file by the customer id.  Include only customers who have invoices.  
    A proc print of
    the merged dataset should look like this (you don't have to use the
    same variable names as I did, but you should get the same values):

  Obs custid custname         custtype   custdate invid    invdate amount

   1   001   Johnson Floral      C     01/02/2002 0003  07/15/2004   5.00
   2   001   Johnson Floral      C     01/02/2002 0008  07/12/2000  16.00
   3   002   Wang Jian           I     01/15/2002 0001  05/15/2004  35.00
   4   002   Wang Jian           I     01/15/2002 0002  06/12/2004  25.50
   5   002   Wang Jian           I     01/15/2002 0004  12/15/2004  72.50
   6   002   Wang Jian           I     01/15/2002 0007  04/12/2005  25.00
   7   003   TrueSale Jewelry    C     06/25/2003 0005  01/25/2005  10.00
   8   005   Jackson Michael     I     12/15/2003 0006  03/15/2005  52.75

5.  Use proc print to display any invoices in which the invoice date is
    earlier in time than the customer contact date.  You should find
    only one such observation:

   Obs custid    custname    custtype   custdate invid    invdate amount

    2   001   Johnson Floral    C     01/02/2002 0008  07/12/2000   16.00

6.  Produce a table showing the total of all invoices for each customer.
    Do not include the invoice with the anomalous date (the one listed
    in part 5).  One possible table format would be the following.  Others
    are fine as long as they include the customer id and the sum and
    have just one row for each customer. 

                   custid      Obs    N             Sum
                   ------------------------------------
                   001           1    1       5.0000000

                   002           4    4     158.0000000

                   003           1    1      10.0000000

                   005           1    1      52.7500000
                   ------------------------------------



7.  Send email to me at

            kcowles@stat.uiowa.edu

    Copy in your entire SAS program and the output from steps 2, 4, 5, 
    and 6.