使用R语言dplyr包处理MLB比赛linescore列：提取各局得分并转换为数值列的技术问询

Hey there! Let's work through your three goals for parsing the MLB linescore data, all using dplyr (plus stringr for regex handling, which plays perfectly with dplyr). Let's start with your existing code and fix/expand it step by step.

1. Extract First Inning Score & Fix Parenthesis Removal #

Your initial code to pull the first inning score was spot-on, but the error with removing parentheses comes because you didn’t wrap str_remove_all inside the mutate() call. Dplyr doesn’t recognize the new inng1 column until the mutate() block finishes, so we need to handle both steps in one go:

library(dplyr)
library(stringr)

# Your original dataset
gamedata <- structure(list(team = c("NYM", "NYM", "BOS", "NYM", "BOS"), linescore = c("010000000", "(10)1140006x", "002200010", "00000(11)01x", "311200"), ondate = structure(c(18475, 18476, 18487, 18489, 18494), class = "Date")), class = "data.frame", row.names = c(NA, -5L))

# Extract first inning and clean parentheses in a single mutate step
gamedata_first_inning <- gamedata %>%
  select(ondate, team, linescore) %>%
  mutate(
    # Grab the first inning: either a single digit or a two-digit value in parentheses
    inng1_raw = str_extract(linescore, "^(\\d|\\(\\d{2}\\))"),
    # Remove parentheses and convert to numeric
    inng1 = as.numeric(str_remove_all(inng1_raw, "[()]"))
  ) %>%
  select(-inng1_raw) # Drop the raw intermediate column if you don't need it

This fixes the "object 'inng1' not found" error because we’re modifying the new column immediately within the same mutate() context.

2. Extract All Subsequent Innings into Numeric Columns #

To parse every inning (including multi-digit scores and trailing x markers), we need to split the linescore into individual inning units, clean them, and reshape the data into wide format. Here’s the full workflow:

gamedata_full_innings <- gamedata %>%
  select(ondate, team, linescore) %>%
  # First, remove the trailing 'x' (marks incomplete 9th inning)
  mutate(linescore_clean = str_remove(linescore, "x$")) %>%
  # Split linescore into individual inning units:
  # Uses regex lookbehind/lookahead to split between valid score chunks
  mutate(innings = str_split(linescore_clean, "(?<=\\d|\\))(?=\\d|\\()")) %>%
  # Unnest the list column to one row per inning per game
  tidyr::unnest_longer(innings) %>%
  # Clean parentheses, assign inning numbers, convert to numeric
  mutate(
    inning_num = row_number(), # Auto-number innings (1, 2, 3...)
    runs = as.numeric(str_remove_all(innings, "[()]"))
  ) %>%
  # Pivot back to wide format with columns like inng1, inng2, etc.
  tidyr::pivot_wider(
    names_from = inning_num,
    values_from = runs,
    names_prefix = "inng"
  )

Key Details on the Regex Split #

The str_split uses a positive lookbehind/lookahead regex ((?<=\\d|\\))(?=\\d|\\()) to correctly split linescores like (10)1140006 into ["(10)", "1", "1", "4", "0", "0", "0", "6"]—it splits right after a digit or closing parenthesis, and right before a digit or opening parenthesis.

Final Output Preview #

Your resulting data will look like this (truncated for clarity):

All inng* columns are numeric, so you can easily calculate total runs with something like rowSums(select(., starts_with("inng")), na.rm = TRUE).

内容的提问来源于stack exchange，提问作者Metsfan