regex find anything which is not a number

myvector <- c("a3", "N.A", "1.2", "-3", "3-2", "2.")
grep('^-?[0-9]+([.][0-9]+)?$',myvector,perl=TRUE,invert=TRUE)

Hi everyone,

I need a regular expression to find those positions in a character
vector which contain something which is not a number (either positive
or negative, having decimals or not).

myvector <- c("a3", "N.A", "1.2", "-3", "3-2", "2.")

In this vector, only positions 3 and 4 are numbers, the rest should be captured.
So far I am able to detect anything which is not a number, excluding - and .

grep("[^-0-9.]", myvector)

[1] 1 2

I still need to capture positions 5 and 6, which in human language
would mean to detect anything which contains a "-" or a "." anywhere
else except at the beginning of a number.

Thanks very much in advance,
Adrian


--
Adrian Dusa
University of Bucharest
Romanian Social Data Archive
Soseaua Panduri nr.90
050663 Bucharest sector 5
Romania

See if the following will work for you:

grep('^-?[0-9]+([.]?[0-9]+)?$',myvector,perl=TRUE,invert=TRUE)

myvector <- c("a3", "N.A", "1.2", "-3", "3-2", "2.")
grep('^-?[0-9]+([.][0-9]+)?$',myvector,perl=TRUE,invert=TRUE)

[1] 1 2 5 6

The key is to match a number, and then invert the TRUE / FALSE (invert=TRUE).
^ == start of string
-? == 0 or 1 minus signs
[0-9]+ == one or more digits

optionally followed by the following via use of (...)?
[.] == an actual period. I tried to escape this, but it failed
[0-9]+ == followed by one or more digits

$ == followed by the end of the string.

so: optional minus, followed by one or more digits, optionally
followed by (a period with one or more ending digits).


On Wed, Mar 11, 2015 at 2:27 PM, Adrian Du?a <dusa.adrian at unibuc.ro> wrote:

Hi everyone,

I need a regular expression to find those positions in a character
vector which contain something which is not a number (either positive
or negative, having decimals or not).

myvector <- c("a3", "N.A", "1.2", "-3", "3-2", "2.")

In this vector, only positions 3 and 4 are numbers, the rest should be captured.
So far I am able to detect anything which is not a number, excluding - and .

grep("[^-0-9.]", myvector)

[1] 1 2

I still need to capture positions 5 and 6, which in human language
would mean to detect anything which contains a "-" or a "." anywhere
else except at the beginning of a number.

Thanks very much in advance,
Adrian


--
Adrian Dusa
University of Bucharest
Romanian Social Data Archive
Soseaua Panduri nr.90
050663 Bucharest sector 5
Romania

--
He's about as useful as a wax frying pan.

10 to the 12th power microphones = 1 Megaphone

Maranatha! <><
John McKown

grep("[^-0-9.]", myvector)

How about letting a standard function decide which are numbers:

which(!is.na(suppressWarnings(as.numeric(myvector))))

Also works with numbers in scientific notation and (presumably) different decimal characters, e.g. comma if that's what the locale uses.

[...]
One problem is that Adrian wanted, for some reason, to exclude numbers
such as "2." but accept "2.0" . That is, no unnecessary trailing
decimal point. as.numeric() will not fail on "2." since that is a
number. The example grep() specifically excludes this by requiring at
least one digit after any decimal point.