在应用函数中使用库函数答案

【问题标题】：Use a library function in the apply function在应用函数中使用库函数
【发布时间】：2019-08-20 12:16:08
【问题描述】：

我有一个数据框，其中包含一个名为“msgText”的列，其中包含文本。为此，我想创建另一个名为“wordcount”的列，用于计算“msgText”每行包含多少个单词。

该列如下所示：

head(all_transcripts$msgText)
[1]  "Hi, my name is Chris and I am a programmer"                                                                                                                                                                                                   
[2]  "I have worked with R for 12 years"                                                                                                                                                                                                                                                                                                                                                           
[3]  "Being a programmer I have many questions" 
[4]  "The fellow programmers at Stackoverflow help me to get the answer"                                                                                                                                                                                  
[5]  "This help has saved my life many times."                                                                                                                                                                                                                                                                                                                                                                                                        
[6]  "Thanks Stackoverflow!"

而我想要的结果是：

head(all_transcripts$wordcount)
    [1]  10                                                                                                                                                                                                   
    [2]  8                                                                                                                                                                                                                                                                                                                                                           
    [3]  7 
    [4]  11                                                                                                                                                                                  
    [5]  8                                                                                                                                                                                                                                                                                                                                                                                                        
    [6]  2

为此，我使用 ngram 库和 wordcount 函数。

我试过了：

all_transcripts$wordcount <- apply(all_transcripts, 2, 
                                   wordcount(all_transcripts$msgText))

但是，这样做我得到了以下错误：

Error in match.fun(FUN) :    'wordcount(all_transcripts$msgText)' is
not a function, character or symbol

如何正确使用 apply 函数，而不必在我的数据集上使用 for 循环？

【问题讨论】：

我认为您正在遍历该列，然后再次提取该列。也许你只需要all_transcripts$wordcount <- wordcount(all_transcripts$msgText)
不行，然后获取整列所有单词的字数
您能发布一小部分数据吗？这样我们就更容易提供帮助了。
Emil，您需要提供一个重现性更高的示例。请以明确格式提供示例数据，例如dput(head(all_transcripts))，以及您的预期/预期输出（因为根据您的描述，您的意思并不清楚）。
apply 的第三个参数中需要一个函数。现在，您得到的不是函数，而是wordcount(all_transcripts$msgText) 返回的任何内容...可能是一个向量

标签： r dataframe apply word-count

【解决方案1】：

考虑向量化的lengths 和strsplit 用于使用base R 进行字数统计：

all_transcripts$word_count <- lengths(strsplit(all_transcripts$text, split=" "))

all_transcripts

#                                                                text word_count
# 1                        Hi, my name is Chris and I am a programmer         10
# 2                                 I have worked with R for 12 years          8
# 3                          Being a programmer I have many questions          7
# 4 The fellow programmers at Stackoverflow help me to get the answer         11
# 5                           This help has saved my life many times.          8
# 6                                             Thanks Stackoverflow!          2

数据

all_transcripts <- data.frame(text=c("Hi, my name is Chris and I am a programmer",
                                     "I have worked with R for 12 years",
                                     "Being a programmer I have many questions",
                                     "The fellow programmers at Stackoverflow help me to get the answer",
                                     "This help has saved my life many times.",
                                     "Thanks Stackoverflow!"),
                              stringsAsFactors=FALSE)

【讨论】：

【解决方案2】：

我们可以遍历“msgText”的元素并应用wordcount函数

library(ngram)
library(tidyverse)
all_transcripts %>%
     mutate(wordcount = map_int(msgText, wordcount))
#                                                             msgText wordcount
#1                        Hi, my name is Chris and I am a programmer        10
#2                                 I have worked with R for 12 years         8
#3                          Being a programmer I have many questions         7
#4 The fellow programmers at Stackoverflow help me to get the answer        11
#5                           This help has saved my life many times.         8
#6                                             Thanks Stackoverflow!         2

或base R

all_transcripts$wordcount <- sapply(all_transcripts$msgText, wordcount)

OP 代码中的问题是它循环通过列（apply 中的MARGIN = 2），其中向量（alltranscripts$wordcount）没有dim 属性

数据

all_transcripts <- structure(list(msgText = c("Hi, my name is Chris and I am a programmer", 
"I have worked with R for 12 years", "Being a programmer I have many questions", 
"The fellow programmers at Stackoverflow help me to get the answer", 
"This help has saved my life many times.", "Thanks Stackoverflow!"
)), class = "data.frame", row.names = c(NA, -6L))

【讨论】：