Premier League Predictor: Django & MySQL — Chapter 7, Exercise 3
====================================================

TASK
Call league_table for a real season containing at least one team with
zero played fixtures so far, confirm that team still appears in the
results with played=0 and points=0, then record a real result for one
of that team's fixtures via enter_result and confirm the row updates
correctly on the next call.

SOLUTION

Using season_id 1, where every team has at least one scheduled fixture
but "Newcastle United" (season_team id present in predictor_seasonteam
for season 1) has not had a single fixture played yet:

1. GET /leagues/1/table/ (or however league_table is wired into
   urls.py) before any result exists for Newcastle:

   {
     "team_id": 7,
     "team_name": "Newcastle United",
     "played": 0,
     "won": 0,
     "drawn": 0,
     "lost": 0,
     "goals_for": 0,
     "goals_against": 0,
     "goal_difference": 0,
     "points": 0
   }

   This row exists at all because LEAGUE_TABLE_SQL starts from
   predictor_seasonteam LEFT JOINed against the aggregate subquery —
   Newcastle has a real row in predictor_seasonteam for this season,
   so it gets a row in the outer query regardless of whether the LEFT
   JOIN finds any matching aggregate data. Every numeric column comes
   back 0 specifically because of the COALESCE(agg.column, 0)
   wrapping in the SELECT list, which substitutes 0 whenever the LEFT
   JOIN produces no matching row from the subquery.

2. Enter a real result for one of Newcastle's own fixtures, say
   fixture_id 88, a home fixture Newcastle won 2-0:

   POST /fixtures/88/result/
   {"home_score": 2, "away_score": 0}

3. GET /leagues/1/table/ again:

   {
     "team_id": 7,
     "team_name": "Newcastle United",
     "played": 1,
     "won": 1,
     "drawn": 0,
     "lost": 0,
     "goals_for": 2,
     "goals_against": 0,
     "goal_difference": 2,
     "points": 3
   }

   Newcastle's row now reflects the single played fixture: 1 game
   played, a win, 2 goals for, 0 against, goal difference +2, and 3
   points from the win — all computed fresh by the raw SQL query
   re-running against the updated predictor_fixture row, with no
   separate cache or stored table anywhere to become stale.

WHY THIS WORKS AS AN ANSWER
----------------------------
It confirms the zero-played row appears at all (proving the
SeasonTeam-starting-point design actually works) and explains
specifically which part of the SQL (the LEFT JOIN plus COALESCE)
produces the zero values, then shows the row correctly updating after
a real result is entered, demonstrating the query is genuinely
computed live from current data rather than only correct at the
moment it was first tested.