Extracting decimal numbers from a string

Question

I have a string such as "3.1 ml" or "abc 3.1 xywazw"

I'd like to extract "3.1" from this string. I have found many questions on stackoverflow about the extraction of numbers from a character string, but no solution works for the case of decimal numbers.

IRTFM · Accepted Answer · 2016-12-20T06:30:00.450

21

This approach makes the decimal point and decimal fraction optional and allows multiple numbers to be extracted:

str <- " test 3.1 test 5"
as.numeric(unlist(regmatches(str,
                             gregexpr("[[:digit:]]+\\.*[[:digit:]]*",str))
          )      )
#[1] 3.1 5.0

The concern about negative numbers can be address with optional perl style look-ahead:

 str <- " test -4.5 3.1 test 5"
    as.numeric(unlist(regmatches(str,gregexpr("(?>-)*[[:digit:]]+\\.*[[:digit:]]*",str, perl=TRUE))))

#[1] -4.5  3.1  5.0

edited Dec 20 '16 at 06:30

answered Oct 08 '13 at 16:55

IRTFM

258,963
21
364
487

Mmm I have found a "bug" today: it drops the minus sign of negative numbers – Stéphane Laurent Feb 27 '14 at 16:38
Put a 'minus' next to the [:digit:] – IRTFM Dec 20 '16 at 06:27
The problem with the `*` quantifier for the decimal point is the potential to match intervals expressed as `1..2` or `2...3` etc. See @Wojciech Sobala's solution with the optional quantifier (`?`). – MilesMcBain Apr 25 '18 at 13:02
If you want to prevent double periods then could also change `*` to `{0,1}` – IRTFM Apr 25 '18 at 15:46

tcash21 · Answer 2 · 2013-10-23T18:25:58.030

20

Use the stringr library:

x<-"abc 3.1 xywazw"
str_extract(x, "\\d+\\.*\\d*")
[1] "3.1"

edited Oct 23 '13 at 18:25

answered Oct 08 '13 at 16:01

tcash21

4,880
4
32
39

8

Possibly use `"\\d+\\.\\d+"` – Ricardo Saporta Oct 08 '13 at 16:03
Thank you. @RicardoSaporta is right. This would not work for `13.1` for instance. – Stéphane Laurent Oct 08 '13 at 17:30
1

And neither would work for "10". I don't think the "+" is the right repetition quantifier to use for the second or third expressions. – IRTFM Oct 08 '13 at 17:38
1

Right @DWin. `str_extract(x, "\\d+\\.*\\d*")` is better. – Stéphane Laurent Oct 16 '13 at 10:11
1

Yes @Stephane, updated the answer to reflect this as it's more generalizeable. – tcash21 Oct 23 '13 at 18:26
Hello tcash21. Do you know how to do to keep the minus sign of negative values ? – Stéphane Laurent Feb 27 '14 at 16:39
I'm using `str_extract(string, "\\-*\\d+\\.*\\d*")`, do you think it is correct ? (it's new hence not very experimented so far) – Stéphane Laurent Feb 27 '14 at 16:47
Try `str_extract(x, "-*\\d+\\.*\\d*")` – tcash21 May 13 '15 at 13:19

score 8 · Answer 3 · answered Oct 08 '13 at 19:25

8

Regular expression for floating point number from http://www.regular-expressions.info/floatingpoint.html with minor adjustment to work in R.

s <- "1e-6 dkel"
regmatches(s,gregexpr("[-+]?[0-9]*\\.?[0-9]+([eE][-+]?[0-9]+)?",s)) 
> [[1]]
> [1] "1e-6"

answered Oct 08 '13 at 19:25

Wojciech Sobala

7,431
2
21
27

Thibaud Ruelle · Answer 4 · 2013-10-08T20:30:33.840

1

You can use regular expressions :

> str <- " test 3.1 test"
> as.numeric(regmatches(str,regexpr("[[:digit:]]+\\.[[:digit:]]+",str)))
[1] 3.1

regexprreturns the start position and length of the matched string. regmatchesreturns the matches. You can then convert it to a number.

edited Oct 08 '13 at 20:30

answered Oct 08 '13 at 16:19

Thibaud Ruelle

303
1
16

My upvote was locked in. I tried to reverse it when I realized the "." was being used inappropriately. It needs to be escaped. Perhaps you can fix that error and the upvote will then be earned. – IRTFM Oct 08 '13 at 16:57
@DWin If you input the text " test 3p1 test" for instance, it is not matched. So I am not sure the "." needs to be escaped here. – Thibaud Ruelle Oct 08 '13 at 19:34
"3p1" is matched, but then converted to NA by `as.numeric`. – IRTFM Oct 08 '13 at 19:42
Please forgive my silly mistake ... I edited my answer. Thanks for noticing. – Thibaud Ruelle Oct 08 '13 at 20:30

score 1 · Answer 5 · answered Jul 25 '23 at 20:12

1

readr::parse_number(c("abc 3.1 xywazw", "-3.1 ml", "1,234.56"))
# [1]    3.10   -3.10 1234.56

answered Jul 25 '23 at 20:12

LMc

12,577
3
31
43

If there is more than one number within a string this only extracts the first instance. – LMc Jul 25 '23 at 20:13

Extracting decimal numbers from a string

5 Answers5

Linked

Related