如何在R语言中基于两个表格创建type-token ratio表格？

步骤1：加载tidyverse包并创建原始数据框 #

先将你提供的形符频率（token）和类符频率（type）数据转换为R数据框：

library(tidyverse)

# 形符频率表
token_df <- tibble(
  AUX = c("toss", "shoot", "launch", "throw", "pull"),
  `13` = c(0, 0, 32, 0, 0),
  `14` = c(0, 0, 15, 0, 0),
  `15` = c(0, 0, 101, 0, 2),
  `16` = c(20, 0, 153, 0, 0),
  `17` = c(47, 3, 95, 0, 0),
  `18` = c(16, 0, 40, 0, 8),
  `19` = c(38, 1, 500, 55, 0),
  `20` = c(11, 1, 500, 125, 7),
  `21` = c(28, 3, 500, 500, 81)
)

# 类符频率表
type_df <- tibble(
  AUX = c("toss", "shoot", "launch", "throw", "pull"),
  `13` = c(0, 0, 8, 0, 0),
  `14` = c(0, 0, 3, 0, 0),
  `15` = c(0, 0, 15, 0, 7),
  `16` = c(16, 0, 12, 0, 0),
  `17` = c(34, 2, 12, 0, 0),
  `18` = c(13, 0, 17, 0, 7),
  `19` = c(32, 1, 18, 44, 0),
  `20` = c(10, 1, 22, 95, 7),
  `21` = c(27, 3, 20, 215, 46)
)

步骤2：转换数据格式并计算比值 #

用pivot_longer将宽格式转为长格式，合并后计算形符-类符比，最后转回宽格式：

# 转换token数据为长格式
token_long <- token_df %>%
  pivot_longer(cols = -AUX, names_to = "year", values_to = "token") %>%
  mutate(year = as.character(year))

# 转换type数据为长格式
type_long <- type_df %>%
  pivot_longer(cols = -AUX, names_to = "year", values_to = "type") %>%
  mutate(year = as.character(year))

# 合并数据并计算type-token ratio
tt_ratio <- token_long %>%
  left_join(type_long, by = c("AUX", "year")) %>%
  # 处理token为0的情况，避免出现Inf/NaN
  mutate(ratio = case_when(
    token == 0 ~ 0,
    TRUE ~ type / token
  )) %>%
  # 保留两位小数
  mutate(ratio = round(ratio, 2)) %>%
  # 转回宽格式
  pivot_wider(id_cols = AUX, names_from = year, values_from = ratio)

步骤3：查看结果 #

运行上述代码后，tt_ratio即为目标表格，打印结果如下：

print(tt_ratio)

内容的提问来源于stack exchange，提问作者Freya