使用逗号分隔字符串从特定数据帧创建频率表

Question

有一个这样的数据帧：

df <- structure(list(doc_id = c("1", "2"), ner_words = c("John, Google", 
"Amazon, Python, Canada")), row.names = c(NA, -2L), class = c("tbl_df", 
"tbl", "data.frame"))

怎么可能做一个table(df$ner_words)，但在每一行中都采用不同的单词？预期结果示例

data.frame(text = c("John", "Google", "Amazon", "Python", "Canada"), frq = c(1,1,1,1,1))

Answer 1

separate_rows()然后count()：

library(dplyr)
library(tidyr)
df %>%
separate_rows(ner_words) %>%
count(ner_words, name = "frq")
# # A tibble: 5 x 2
#   ner_words   frq
#   <chr>     <int>
# 1 Amazon        1
# 2 Canada        1
# 3 Google        1
# 4 John          1
# 5 Python        1

Answer 2

这是一个选项：

library(dplyr)
library(tidyr)
df %>% 
separate_rows(ner_words, sep = ", ") %>% 
group_by(ner_words) %>% 
transmute(ner_words, frq = n())
# A tibble: 5 x 2
# Groups:   ner_words [5]
ner_words   frq
<chr>     <int>
1 John          1
2 Google        1
3 Amazon        1
4 Python        1
5 Canada        1

Answer 3

带有str_extract和count的选项

library(dplyr)
library(stringr)
library(tidyr)
df %>%
transmute(ner_words = str_extract_all(ner_words, '\w+')) %>% 
unnest(c(ner_words)) %>%
count(ner_words)
# A tibble: 5 x 2
#  ner_words     n
#* <chr>     <int>
#1 Amazon        1
#2 Canada        1
#3 Google        1
#4 John          1
#5 Python        1