R 变量的 Grep 文件名

Question

I am new to R so I am struggling with what I imagine is a fairly simple question. For this question I am not looking for someone to give me just a solution. I was hoping that someone could explain the answer to me, so that I might learn to do it myself, rather than just copy what it is you have done. That being said, here is my problem and questions.

I am making a histogram with R. A user will submit a file and data from that file will be used to make a histogram. That much is already set and done. Where I am having a problem is that I need to take only part of that file name and use it to help make a title for the histogram. The file name is a bit of a monster and follows this naming convention:

X_Y.doc.Z.x_y_z

The aspects of that file name that I need are the Y and Z. I know that many people use grep but I am not sure how to use it in this instance. I have already read the ??grep page and am familiar with the basics of grep but don't really know where to start.

Eventually I will also need to grep some information from an excel file, if someone cares to advise me in that matter as well. If it helps, this is how I am accepting the files:

F.n<-(tk_choose.files(default="", caption="Select a file", multi=TRUE, filters=NULL, index=1))

Does anyone have any suggestions?

Answer 1

The answer already given using stringr is excellent. That package provides you with some very helpful string munging tools.

If you want to only use base, you could do this with gsub . Assuming your punctuation stays the same and there will not be any embedded periods or underscores in the X, Y or Z something like this should work

f <- 'X_Y.doc.Z.x_y_z'
gsub('^.+_(.+)\\.doc\\.(.+)\\..+_.+$', '\\1 \\2', f)

which returns:

"Y Z" 

you could put whatever you want in there though to make it easier to get at each piece or could do this in two lines returning one each. And remember, R almost never changes data in place. You need to assign the output of a function to a variable like below. Otherwise it will just print to the console and be "lost" (this is true most of the time).

y <- gsub('^.+_(.+)\\.doc\\..+\\..+_.+$', '\\1', f)
z <- gsub('^.+_.+\\.doc\\.(.+)\\..+_.+$', '\\1', f)

Lets break it down.

^ specifies the beginning of a line. its good to be explicit. similarly $ identifies the end of a line.

. represents any character and following it with a + means one or more of any character. If you used .* instead of .+ it would mean zero or more of any character and that isnt what we want. If i want to write a normal . I need to escape it since its a special character. \\ is the escape character both for regular expressions and for R. So... you need two. To write a normal period you need to write \\\\.

Clear to be sure. Finally the parentheses represent a group I want to save. They can be referenced later using numbers indicating the order you saved them. In some languages these parentheses need to be escaped also, but not R.

Answer 2

Grep uses Regular Expressions to search for substrings matching a pattern. For your problem of matching certain elements from a filename, you would probably want to use capturing groups to extract the different parts.

An example of a regular expression with a capturing group would be:

"Hello, (\w+)"

To match strings of the format "Hello, Friend". Here is an explanation of the pattern:

• \\w will match a "word character", while
• + means that at least one, but multiple of them will be matched.
• For the other structural parts of your file name convention, we can just include _ as they are but have to escape . as they have a special meaning in regular expressions.
• To define a group that you want to match (a capturing group), you put the part to be matched in parentheses (\\w+)

Using all that, we get the following pattern:

"(\w+)_(\w+)\.doc\.(\w+)\.(\w+)_(\w+)_(\w+)"

To get the pattern to work in R, we will have to escape all \\ characters as \\\\ :

> pattern = "(\\w+)_(\\w+)\\.doc\.(\\w+)\\.(\\w+)_(\\w+)_(\\w+)"

While grep and regex are powerful, I personally prefer the stringr package for its simpler interface, in particular the str_match function can be very helpful as it will return a matrix with column 1 giving the full match and all subsequent columns giving the matches to the capturing groups:

> x = "X_Y.doc.Z.x_y_z"
> str_match(x, pattern)

     [,1]              [,2] [,3] [,4] [,5] [,6] [,7]
[1,] "X_Y.doc.Z.x_y_z" "X"  "Y"  "Z"  "x"  "y"  "z" 

If you are new to regular expressions, you should be fine with a tutorial for any language such as this one . Syntax will mostly be similar, but vary only in details while not all features are supported by all programming languages. If you want to try out your expressions before putting them into your programs, I highly recommend RegexPal

Answer 3

In this simple case of just needing a single letter that is in a well-defined place, substr would probably be simpler:

> a <- "X_Y.doc.Z.x_y_z"
> substr(a, 3, 3)
[1] "Y"
> substr(a, 9, 9)
[1] "Z"