Detect alpha + digit + dash : must contain numeric

Question

I want to logically, find IDs in a file. These will have only digits, letters, and dashes. They must containa digit to be considered. I could do a Boolean with 2 grepl statements but want to do this with a single regex. I think (SKIP)(FAIL) could work but don't know how. In the following I want elements 1, 2, 5, 6 to be considered IDs.

g <- c(
    "868776767-ddd-dFFF-999999",
    "8888888",
    "bbbbbbfdfdgtfref-dsfcsdbcgwecbgfecshdcs-cdhscgbfsd",
    "bigbird",
    "2",
    "3-4",
    "swe%h"
)

## This works (I want this result with one regex)
grepl("[A-Za-z0-9-]+", g) & grepl("[0-9]+", g)

## I suspect using this could work with a single regex call.
grepl("(*SKIP)(*FAIL)", g)

Check this article http://stackoverflow.com/questions/469913/regular-expressions-is-there-an-and-operator — Mariano, Sep 16 '15 at 17:32

Casimir et Hippolyte · Accepted Answer · 2015-09-16T19:18:03.027

No need to search complicated things:

grepl("^[a-fA-F-]*[0-9][[:xdigit:]-]*$", g)

or

grepl("^[a-fA-F-]*+[[:xdigit:]-]+$", g, perl=T)

where [:xdigit:] is the POSIX character class that contains [a-fA-F0-9]. The second version uses a possessive quantifier to ensure that the next character is a digit.

If you want to ensure that there is no leading, trailing or consecutive hyphens:

grepl("^(?:[a-fA-F]+(?:-[a-fA-F]+)*)?[0-9][[:xdigit:]]*(?:-[[:xdigit:]]+)*$", g)

score 0 · Answer 2 · edited Sep 16 '15 at 23:47

0

You can use the following:

^(?=.*\d)[a-zA-Z0-9-]*$

Explanation:

^ : start of the string
(?=.*\d) : look ahead for atleast one digit
[a-zA-Z0-9-]+ : match more than one alpha or digit or -
$ : end of the string

Output:

grepl("^(?=.*\\d)[a-zA-Z0-9-]*$", g, perl=TRUE)
## [1]  TRUE  TRUE FALSE FALSE  TRUE  TRUE FALSE

edited Sep 16 '15 at 23:47

Tyler Rinker

108,132
65
322
519

answered Sep 16 '15 at 17:32

karthik manchala

13,492
1
31
55

@TylerRinker no problem.. i think now it will help people working with r too :) – karthik manchala Sep 16 '15 at 17:42
Why add the wildcard in the lookaround? `(?=.*\\d)` – Pierre L Sep 16 '15 at 17:47
@PierreLafortune because the string could have only one digit and that too at the end.. – karthik manchala Sep 16 '15 at 17:49
Yes but according to the conditions one digit is okay. I meant specifically `.*`. – Pierre L Sep 16 '15 at 17:51
@PierreLafortune yes.. and how would you ensure atleast one digit? – karthik manchala Sep 16 '15 at 17:52
`grepl("^(?=\\d)[a-zA-Z0-9-]+$", g, perl=TRUE)` – Pierre L Sep 16 '15 at 17:53
@PierreLafortune that regex will check for a digit at the beginning only.. try matching `aaa-123` with it and see what happens.. :) – karthik manchala Sep 16 '15 at 17:54
Should the plus quantifier be. `*` to match just one digit? – Tyler Rinker Sep 16 '15 at 23:10
@TylerRinker `*` means zero or more letters and `+` means one or more. Here we have an additional assertion for atleast one digit, so both will give same results always :) – karthik manchala Sep 17 '15 at 03:10

Detect alpha + digit + dash : must contain numeric

2 Answers2