From c9886313b5a6a895677b2e5a414ede769678d69e Mon Sep 17 00:00:00 2001
From: RGAlexander216 <20406603+RGAlexander216@users.noreply.github.com>
Date: Sun, 15 Sep 2024 20:24:34 -0500
Subject: [PATCH] Fix pbp and pbp participation join issue (#79)

Resolves issue with play_by_play data having incorrect old_game_id values. For example, all old_game_id values in the play_by_play_2023 for week 15 start with "2022" but they begin with "2023" in the pbp_participation_2023 data.
---
 nfl_data_py/__init__.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/nfl_data_py/__init__.py b/nfl_data_py/__init__.py
index c6a2ecc..bdacfba 100644
--- a/nfl_data_py/__init__.py
+++ b/nfl_data_py/__init__.py
@@ -146,7 +146,10 @@ def import_pbp_data(
                 if all([include_participation, year >= 2016, not cache]):
                     path = r'https://github.com/nflverse/nflverse-data/releases/download/pbp_participation/pbp_participation_{}.parquet'.format(year)
                     partic = pandas.read_parquet(path)
-                    raw = raw.merge(partic, how='left', on=['play_id','old_game_id'])
+                    raw = raw.merge(partic,
+                                    how='left',
+                                    left_on=['play_id','game_id'],
+                                    right_on=['play_id','nflverse_game_id'])
                 
                 pbp_data.append(raw)
                 print(str(year) + ' done.')