NBA_Tutorials icon indicating copy to clipboard operation
NBA_Tutorials copied to clipboard

Handle revisions in play-by-play data

Open Brandon-Michaud opened this issue 1 year ago • 1 comments

I noticed that sometimes the play-by-play csv files have revisions when an event was forgotten and it was added at a later time. This means the event has a higher EVENTNUM, but the PCTIMESTRING displays the actual time of the event. This commit changes the starters parser to handle such revisions, whereas it could not in the past.

I noticed one issue: When a player is subbed in and out during the same period and time on PCTIMESTRING and the first event was missed and included as a revision. This could be being subbed in before a free throw, the free throw was made, a timeout was called, and being subbed out. This is an incredibly rare event and the only game IDs I found it being an issue are 0029701045 0020500090 0022201040

Brandon-Michaud avatar Jun 17 '24 21:06 Brandon-Michaud

Before these changes, the parser would find a number of starters other than 5, creating issues for the next step of parsing possessions

Brandon-Michaud avatar Jun 17 '24 21:06 Brandon-Michaud