CODEWITHSUNDEEP
regexr
BRZ
BRZ

Reputation: 695

Grep for Pattern in File in R

In a document, I'm trying to look for occurences of a 12-digit string which contains alpha and numerals. A sample string is: "PXB111X2206"

I'm trying to get the line numbers that contain this string in R using the below:

FileInput = readLines("File.txt") 
prot_pattern="([A-Z0-9]{12})";
prot_string<-grep(prot_pattern,FileInput)
prot_string

This worked fine until it hit a document containing all upper-case titles and returned a line containing the word "CONCENTRATIO"

The string I am trying to look for is: "PXB111X2206". I am expecting the grep to return the line numbers containing the string : "PXB111X2206". It however is returning the line number containing the word: "CONCENTRATIO"

What is wrong with my expression above? Any idea what I am doing wrong here?

Here is some sample input: Each design objective described herein is significantly important, yet it is just one aspect of what it takes to achieve a successful project.

A successful project is one where project goals are identified early on and where the >interdependencies of all building systems are coordinated concurrently from the planning and programming phase.

CONCENTRATION:

The areas of concentration for design objectives: accessible, aesthetics, cost effective, >functional/operational, historic preservation, productive, secure/safe, and sustainable and >their interrelationships must be understood, evaluated, and appropriately applied.

Each of these design objectives is presented in the design objectives document number. >PXB111X2206. >

Thanks & Regards, Simak

Upvotes: 0

Views: 17151

Answers (2)

amaurs
amaurs

Reputation: 1642

You are using a very powerful tool for a very simple task, the expression 

[A-Z0-9]{12}

will match any alphanumeric 12 sized uppercased string, for example the word "CONCENTRATIO", however, your "PXB111X2206" is not even 12 symbols long, so it is not possible that is being matched. If you only want to match "PXB111X2206" you only have to use it as a regular expression itself, for example, if you file contents are:

foo
CONCENTRATIO. 
bazz
foo bar bazz PXB111X2206 foo bar bazz
foo 
bar
bazz

and you use:

grep('PXB111X2206',readLines("File.txt"))

then R will only match line 4 as you would wish.

EDIT

If you are looking for that specific pattern try:

grep('[A-Z]{3}[0-9]{3}[A-Z]{1}[0-9]{4}',readLines("File.txt"))

That expression will match strings like 'AAADDDADDDD' where A is an capital letter, and D a digit, the regular expression contains a group (symbols inside square brackets) and a quantifier (the number inside the brackets) that tells how many of the previous symbol will the expression accept, if no quantifier is present it assumes it is 1.

Upvotes: 3

nwknoblauch
nwknoblauch

Reputation: 568

Let's take a look at what your regular expression means. [A-Z0-9] means any capitalized letter or number and {12} means the previous expression must occur exactly 12 times. The string CONCENTRATIO is 12 capitaized letters, so it's no surprise that grep picks it up. If you want to take out the matches that match to just letters or just numbers you could try something like 

allleters <- grep("[A-Z]{12}",strings)
allnumbers <-grep("[0-9]{12}",strings)
both <- grep("[A-Z0-9]{12}",strings)

the matches you wanted would then be something like 

both <- both[!both %in% union(allletters,allnumbers)]

Someone with better regexfu might have a more elegant solution, but this will work too.

Upvotes: 0

Related Questions