Aggregate function not working correctly

Question

I have this data in a table
SZSTCLA_PIDM     SZSTCLA_TERM_CODE       SZSTCLA_TERM_CODE_MATRIC     SZSTCLA_TOTAL_CREDITS
82724                        201010                                               201010                      4.50
82724                       201020                                              201010                      4.50
82724                       201110                                              201010                      4.50
82724                       201120                                              201010                      4.00
82724                       201210                                              201010                      0.00
82724                      201220                                              201010                     0.00
82724                      201310                                              201010                     3.00 when I run this query
SELECT szstcla_pidm,
                  szstcla_term_code,
                  SUM (
                     SZSTCLA_TOTAL_CREDITS)
                  OVER (PARTITION BY szstcla_pidm
                        ORDER BY szstcla_term_code
                        ROWS BETWEEN UNBOUNDED PRECEDING AND 1 PRECEDING)
                     AS aggr_hours,
                  szstcla_term_code AS szstcla_term_code_2
             FROM szstcla
            WHERE                                  ---szstcla_pidm    =  82085
                 szstcla_term_code >= szstcla_term_code_matric
                 AND SZSTCLA_TOTAL_CREDITS > 0
         ORDER BY szstcla_pidm I ended with
82724     201010          201010
82724     201020     4.5     201020
82724     201110     9     201110
82724     201120     13.5     201120
82724     201310     17.5     201310 It is not counting
82724                      201310                                              201010                     3.00 I should ended with
82724     201010          201010
82724     201020     4.5     201020
82724     201110     9     201110
82724     201120     13.5     201120
82724     201310     17.5     201310
82724     201320     20.5     201320

Answer 1

Chris and Karthic have it right, there is no 201320 term code so you won't get a record for that term until one is inserted in your table and your sums are based on the window you specified.
What I'm wondering though is if you want to partition by the SZSTCLA_TERM_CODE_MATRIC column as well as SZSTCLA_PIDM. In your where clause you are limiting the rows to those where SZSTCLA_TERM_CODE is >= to SZSTCLA_TERM_CODE_MATRIC, but are you likely to have rows with SZSTCLA_TERM_CODE < SZSTCLA_TERM_CODE_MATRIC? There weren't any in your sample data, but I would expect that you may have multiple matriculation terms for a given PIDM. If you want to restart the calculation every time the student matriculates, you'll want to partition by that matriculation term.
Have a look:
with szstcla as (
  select 82724 SZSTCLA_PIDM, '201010' SZSTCLA_TERM_CODE, '201010' SZSTCLA_TERM_CODE_MATRIC
         , 4.50 SZSTCLA_TOTAL_CREDITS from dual union all
  select 82724, '201020', '201010', 4.50 from dual union all
  select 82724, '201110', '201010', 4.50 from dual union all
  select 82724, '201120', '201120', 4.00 from dual union all
  select 82724, '201210', '201120', 0.00 from dual union all
  select 82724, '201220', '201120', 0.00 from dual union all
  select 82724, '201310', '201120', 3.00 from dual
select szstcla_pidm PIDM
     , szstcla_term_code_matric TC_MATRIC
     , szstcla_term_code TERM_CODE
     , sum (szstcla_total_credits)
       over (partition by szstcla_pidm
             order by szstcla_term_code
             rows between unbounded preceding and 1 preceding) agg_hours
     , sum (szstcla_total_credits)
       over (partition by szstcla_pidm
             order by szstcla_term_code
             rows between unbounded preceding and current row) agg_hours2
     , sum (szstcla_total_credits)
       over (partition by szstcla_pidm, szstcla_term_code_matric
             order by szstcla_term_code
             rows between unbounded preceding and 1 preceding) agg_hours3
     , sum (szstcla_total_credits)
       over (partition by szstcla_pidm, szstcla_term_code_matric
             order by szstcla_term_code
             rows between unbounded preceding and current row) agg_hours4
  from szstcla
  where szstcla_term_code >= szstcla_term_code_matric
    AND SZSTCLA_TOTAL_CREDITS > 0
  order by szstcla_pidm, szstcla_term_code;
PIDM     TC_MATRIC     TERM_CODE     AGG_HOURS     AGG_HOURS2     AGG_HOURS3     AGG_HOURS4
82724     201010          201010                    4.5                    4.5
82724     201010          201020          4.5          9          4.5          9
82724     201010          201110          9          13.5          9          13.5
82724     201120          201120          13.5          17.5                    4
82724     201120          201310          17.5          20.5          4          7I modified your original sample data to include a change in matriculation term codes part way through the data. The last AGG_HOURS and AGG_HOURS3 share the same windowing clause as your original query, AGG_HOURS2 and AGG_HOURS4 share the windowing clause proposed by Chris, and AGG_HOURS3 and AGG_HOURS4 add the additional partition on the matriculation term code.