英文:
r doesn't see data frame as characters; can't grep; wrong use of as.character()?
问题
编辑，一天后：对这个问题的答案告诉我我需要相当大幅度地编辑我的代码。所以基本上这个问题现在已经不存在了，因为我不再在数据框中使用grep了。现在的代码如下，更加清晰。
我将原始问题保留在这里，以防我的学习过程能帮助任何人，尽管如此。
# 1. Find lines containing both "un" and "ɛ̃"original_lines <- readLines('Test.txt')lines_with_pattern <- grep('un.*ɛ̃', original_lines, value = TRUE)# CHANGE PHONES TO FIND AND PHONES TO ADD# 2. Duplicate the line in which the pattern occurs and change the relevant phonememodified_lines <- character()for (line in lines_with_pattern)modified_lines <- c(modified_lines, gsub("ɛ̃", "œ̃", line))# 3. Combine modified lines with original linesall_lines <- c(original_lines, modified_lines)# 4. Sort the lines alphabeticallysorted_lines <- sort(all_lines)# 5. Print the sorted lineswriteLines(sorted_lines, 'myfile.txt', sep = '\n')
原始问题：
我试图在一个由两列组成的数据框中使用grep，两列之间以制表符分隔，例如：
V1 V217 nempruntèrent ɑ̃ p ʁ ɛ̃ t ɛ ʁ18 vemprunté ɑ̃ p ʁ ɛ̃ t e19 fempruntée ɑ̃ p ʁ ɛ̃ t e20 wempruntées ɑ̃ p ʁ ɛ̃ t e21 2empruntés ɑ̃ p ʁ ɛ̃ t e
（摘录--数据框的最后五行。第一列包含类似法语的虚构单词；第二列包含国际音标字母表的音标。）
test <- read.delim('Test.txt', header=FALSE)print(test)
产生如上所示的输出，所以看起来R“知道”数据框中有什么。
但是然后我想要在其中使用grep来查找特定的字符串，所以我尝试了以下操作：
# 1. Find lines containing both "un" and "ɛ̃"lines_with_pattern <- grep("un", test, value = TRUE)print(lines_with_pattern)
这不起作用。
上面的grep结果是named character(0)。这意味着R找不到它正在寻找的字符，因此我尝试了以下操作：
test <- read.delim('Test.txt', header=FALSE)test <- as.character(test)
我认为我没有正确使用as.character()，因为这段代码产生了例如：
V1 V2[17,] NA NA[18,] NA NA[19,] NA NA[20,] NA NA[21,] NA NA
（再次是输出的最后五行）
因此print(test)产生了：
[1] "c(17, 18, 12, 20, 1)"[2] "c(8, 6, 6, 6, 6)"
（结果向量中的最后五个数字）
和
lines_with_pattern <- grep("un", test, value = TRUE)# value = TRUE, fixed = FALSE, useBytes = TRUE, invert = FALSE)print(lines_with_pattern)
产生了character(0)。
所以：我不理解print(test)产生的向量，上面的例子中的数字似乎与数据对应的任何内容都不符。而且，我的原始问题是：我需要做什么才能在这个数据集中使用grep？
抱歉信息很长，也抱歉提了一个初学者的问题，但非常感谢任何帮助！
英文:
EDIT, A DAY LATER: The answers to this question have shown me that I needed to edit my code quite substantially. So basically the question has now gone away because I'm not grepping in a data-frame any more. The code is now as follows, much cleaner.
I'm leaving the original question here in case my learning process helps anyone, though.
# 1. Find lines containing both "un" and "ɛ̃"original_lines <- readLines('Test.txt')lines_with_pattern <- grep('un.*ɛ̃', original_lines, value = TRUE)# CHANGE PHONES TO FIND AND PHONES TO ADD# 2. Duplicate the line in which the pattern occurs and change the relevant phonememodified_lines <- character()for (line in lines_with_pattern)modified_lines <- c(modified_lines, gsub("ɛ̃", "œ̃", line))# 3. Combine modified lines with original linesall_lines <- c(original_lines, modified_lines)# 4. Sort the lines alphabeticallysorted_lines <- sort(all_lines)# 5. Print the sorted lineswriteLines(sorted_lines, 'myfile.txt', sep = '\\n')
ORIGINAL QUESTION
I am trying to grep a data frame consisting of rows of two columns with tab separation between the columns, e.g.
V1 V217 nempruntèrent ɑ̃ p ʁ ɛ̃ t ɛ ʁ18 vemprunté ɑ̃ p ʁ ɛ̃ t e19 fempruntée ɑ̃ p ʁ ɛ̃ t e20 wempruntées ɑ̃ p ʁ ɛ̃ t e21 2empruntés ɑ̃ p ʁ ɛ̃ t e
(excerpt--the last five rows of the data frame. The first col contains dummy French-like words; the second col contains International Phonetic Alphabet transcriptions.)
test <- read.delim('Test.txt', header=FALSE)print(test)
produces a printout as above, so it looks as if R 'knows' what's in the data frame.
But then I want to grep for certain strings, so I've tried
# 1. Find lines containing both "un" and "ɛ̃"lines_with_pattern <- grep("un", test, value = TRUE)print(lines_with_pattern)
and this doesn't work.
The above grep results in named character(0). This means R is not finding the characters it's looking for, so I've tried
test <- read.delim('Test.txt', header=FALSE)test <- as.character(test)
I don't think I'm using as.character() right, as that snippet produces e.g.
V1 V2[17,] NA NA[18,] NA NA[19,] NA NA[20,] NA NA[21,] NA NA
(again the last five rows of the output)
and therefore print(test) produces
[1] "c(17, 18, 12, 20, 1)"[2] "c(8, 6, 6, 6, 6)"
(last five figures in the resulting vectors)
and
lines_with_pattern <- grep("un", test, value = TRUE)# value = TRUE, fixed = FALSE, useBytes = TRUE, invert = FALSE)print(lines_with_pattern)
produces character(0).
So: I don't understand the vectors that print(test) produces in the example just above--the numbers don't seem to refer to anything corresponding to the data. And, my original question: what do I need to do to be able to grep this data-set?
Sorry for the very long message, and for the noob question, but thanks a lot for any help!
答案1
得分: 3
`grep()` 不能直接用于数据框。此外，数据框在逐行操作方面效果不佳，这似乎是您感兴趣的操作。在基础 R 中，我会使用 `apply()`（它会隐式将数据转换为字符矩阵）和 `grepl()` 的组合：```rwhich(apply(test,MARGIN = 1,FUN = function(x) any(grepl("un", x)) & any(grepl("ɛ̃", x))))
这将给您所有包含 "un" 和 "ɛ̃" 的行。使用 which() 是为了获得行号，而不是逻辑向量（这也非常适用于单独进行子集选择）。
<details><summary>英文:</summary>`grep()` cannot be used as-is on data frames. Moreover, data frame's do not work well for row-wise operations which is what you seem to be interested in doing. In base R, I would use a combination of `apply()` (which will do an implicit conversion of your data to a character matrix) and `grepl()`:```rwhich(apply(test,MARGIN = 1,FUN = function(x) any(grepl("un", x)) & any(grepl("ɛ̃", x))))
This will give you all the rows where both "un" and "ɛ̃" appear. The use of which() is to get row numbers rather than a logical vector (which also works perfectly fine for subsetting on its own).
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论