R subset/keep 至少包含两个特定文本字符串的所有行

Question

我有一个包含不同文本摘录的数据框。

我希望对包含我的小词典（“贫困|报告|警报|inflation”）中至少 2 个术语 的所有观察结果进行子集化，或者同一个术语两次（例如 report 在文本中出现两次）。

texts <- data.frame(text = c("report highlights that poverty is widespread", "there is inflation", "alarming reports", "thanks for listening"), id = 1:4, group = 4:7)

texts[grepl("poverty|report|alarming|inflation", texts$text, ignore.case=T),]
#      I don't want this:                      text id group
#1 report highlights that poverty is widespread  1     4
#2                           there is inflation  2     5
#3                             alarming reports  3     6

但我想要这个：

#                                                     text id group
#1 report highlights that poverty is widespread         1     4
#3                         alarming reports             3     6

Answer 1

这个有用吗：

> library(stringr)
> library(dplyr)
> texts %>% filter(str_count(text, pattern = "poverty|report|alarming|inflation") > 1)
                                          text id group
1 report highlights that poverty is widespread  1     4
2                             alarming reports  3     6
>

Answer 2

尝试这种base R方法：

#Data
texts <- data.frame(text = c("report highlights that poverty is widespread", "there is inflation", "alarming reports", "thanks for listening"), id = 1:4, group = 4:7,stringsAsFactors = F)
#Index
Index <- apply(texts[,1,drop=F],1,function(x)sum(grepl("poverty|report|alarming|inflation",
                                                       unlist(strsplit(x,split =' ')),
                                                       ignore.case=T)))
#Subset
texts[which(Index>=2),]

输出：

                                          text id group
1 report highlights that poverty is widespread  1     4
3                             alarming reports  3     6

R subset/keep 至少包含两个特定文本字符串的所有行

R subset/keep all rows with at least two specific text strings

text

r

grepl

tidyr