Function for finding all perfect-k games from Retrosheet data for a particular season

perfect.R

perfect <- function(d, n_innings){
  # inputs are Retrosheet play-by-play data frame d and
  # number of innings n_innings
  # output is a data frame containing all Perfect-k
  # games for that season
  
  require(dplyr)
  
  d %>% 
    mutate(PIT_ID = as.character(PIT_ID),
           GAME_ID = as.character(GAME_ID)) -> d
  
  # N_opp <- 2 * length(unique(d$GAME_ID))
  
  # first consider top of inning (BAT_HOME_ID = 0)
  
  d %>% 
    filter(BAT_HOME_ID == 0,
           BAT_EVENT_FL == TRUE) %>% 
    group_by(GAME_ID, INN_CT) %>% 
    summarize(N = n(),
              S = sum(EVENT_CD %in% c(2, 3)),
              Perfect = ifelse(N == S, 1, 0),
              PIT_ID_first = first(PIT_ID),
              Pit_N = sum(PIT_ID == PIT_ID_first),
              .groups = "drop") -> 
    S_top
  
  # now bottom of inning
  
  d %>% 
    filter(BAT_HOME_ID == 1,
           BAT_EVENT_FL == TRUE) %>% 
    group_by(GAME_ID, INN_CT) %>% 
    summarize(N = n(),
              S = sum(EVENT_CD %in% c(2, 3)),
              Perfect = ifelse(N == S, 1, 0),
              PIT_ID_first = first(PIT_ID),
              Pit_N = sum(PIT_ID == PIT_ID_first),
              .groups = "drop") -> 
    S_bottom
  
  # find pitchers with n_innings perfect innings
  
  S_top %>% 
    filter(INN_CT <= n_innings) %>% 
    group_by(GAME_ID) %>% 
    summarize(Pitcher_First = first(PIT_ID_first),
              Pitcher_Last = last(PIT_ID_first),
              N_Perfect = sum(Perfect),
              N_Pitcher = sum(Pit_N),
              .groups = "drop") %>% 
    filter(Pitcher_First == Pitcher_Last,
           N_Perfect == n_innings,
           N_Pitcher == 3 * n_innings) %>% 
    select(GAME_ID, Pitcher_First) -> S2_top
  
  S_bottom %>% 
    filter(INN_CT <= n_innings) %>% 
    group_by(GAME_ID) %>% 
    summarize(Pitcher_First = first(PIT_ID_first),
              Pitcher_Last = last(PIT_ID_first),
              N_Perfect = sum(Perfect),
              N_Pitcher = sum(Pit_N),
              .groups = "drop") %>% 
    filter(Pitcher_First == Pitcher_Last,
           N_Perfect == n_innings,
           N_Pitcher == 3 * n_innings) %>% 
    select(GAME_ID, Pitcher_First) -> S2_bottom
  
   P_out <- rbind(S2_top, S2_bottom)
   P_out$N_Innings <- n_innings
   
   P_out$Season <- substr(d$GAME_ID[1], 4, 7)
  
   P_out
}