Pattern matching digits does not work in egrep?

Question

Why can't I match the string

"1234567-1234567890"

with the given regular expression

\d{7}-\d{10}

with egrep from the shell like this:

egrep \d{7}-\d{10} file

?

Marquise

I just tried /\d{7}-\d{10}/ and it works fine with that string. – Burnsed

It doesn't work; I've written that string from above inside file - but nothing ?! – Marquise

@persistent: maybe you need to whip up a hex editor and see what those digit characters really are. – Potion

Hey :D ; just a simple digits forwarded from the standard output with echo; echo string > file – Marquise

Answer 1

egrep doesn't recognize \d shorthand for digit character class, so you need to use e.g. [0-9].

Moreover, while it's not absolutely necessary in this case, it's good habit to quote the regex to prevent misinterpretation by the shell. Thus, something like this should work:

egrep '[0-9]{7}-[0-9]{10}' file

See also

• egrep mini tutorial

References

• regular-expressions.info/Flavor comparison
  • Flavor note for GNU grep, ed, sed, egrep, awk, emacs
    • Lists the differences between grep vs egrep vs other regex flavors

Potion

Actually he only needs to quote the regex if it contains shell meta-characters. And now that it no longer contains backslashes, it doesn't, so quoting is optional. – Grecian

@sepp2k: do you need quote for a space? I think you do. I guess you can argue that a space is a shell metacharacter. Anyway I think it's best to always quote, ala it's best to always use curly braces. – Potion

Then how would be with grep instead; I'm interested in \d prefix?! – Marquise

@persistent: according to comparison chart I linked, neither POSIX ERE (egrep) nor POSIX BRE (grep) knows \d, \s, \w, \b, etc. Also \d is not a prefix; it's a shorthand for the digit character class supported by many but not all flavors. – Potion

Well that's odd; then where they're specified if not inside (e)grep ? – Marquise

It's not prefix; inside the sintax it must be specified as a prefix before the brackets; thx for the correction anyway ;) – Marquise

@persistent: different flavors of regex does things differently, that's why it's important to mention which flavor you're using when asking regex questions, etc. I'll guess that Perl popularized the \d shorthand, and everyone else followed later. – Potion

@polygenelubricants: Yes, you need quotes with spaces (or put a backslash before every space). And sure, it doesn't hurt to always quote. – Grecian

@persistent: you can't use \d with grep/egrep; you can use its expanded form [0-9] which is practically the same thing, but slightly longer. In some flavors that supports Unicode, \d is not the same as [0-9] because it also includes some other Unicode digit characters. – Potion

Well mate; I already knew for the [block]{} form ; I was interested in \d; thx – Marquise

I don't ever use [0-9], when I really mean [[:digit:]]. Plus these character classes are supported almost everywhere, and are defined in POSIX. – Tisbe

'\d' wasn't working for me, but '\s' and '\w' did today. grep (GNU grep) 2.23 Packaged by Homebrew. – Feline

Answer 2

For completeness:

Egrep does in fact have support for character classes. The classes are:

• [:alnum:]
• [:alpha:]
• [:cntrl:]
• [:digit:]
• [:graph:]
• [:lower:]
• [:print:]
• [:punct:]
• [:space:]
• [:upper:]
• [:xdigit:]

Example (note the double brackets):

egrep '[[:digit:]]{7}-[[:digit:]]{10}' file

Bildungsroman

Just a complaint about grep: [[:digit:]] is worse than [[0-9]] in every possible way. None of these are short hand, and they are harder to rememer than the default regex syntax. EG: [[:lower:]] is harder to remember, read and write than [a-z] – Godber

@Godber But grep -E i.e. egrep supports both [:digit:] and [0-9] so where is the complaint? If you're comparing with \d it's arguable. d could stand for anything, a bit like one letter variable names. Still seems \d has become more popular. I think grep character classes pre-date Perl \d – Immolation

Answer 3

you can use \d if you pass grep the "perl regex" option, ex:

grep -P "\d{9}"

Bhatt

Yes. Thank you! – Chock

Not in FreeBSD. Option -P is not supported, unfortunately. – Nowell

If you're running AWS CodeBuild you'll still have this issue and you'll need to use grep -P – Dunfermline

Answer 4

Use [0-9] instead of \d. egrep doesn't know \d.

Grecian

According to regular-expressions.info/gnu.html, repetition in grep is \{7\}. – Potion

@polygenelubricants: Sure, I thought he was asking about \d. – Grecian

Answer 5

try this one:

egrep '(\d{7}-\d{10})' file

Manriquez

Traditional egrep did not support the { metacharacter, and some egrep implementations support \{ instead, so portable scripts should avoid { in egrep patterns and should use [{] to match a literal {. – Manriquez

However neither traditional egrep nor GNU egrep support \d and that's why this does not work - not because of the {. Though it'd be useful to keep the { thing in mind if you ever have to be compatible with traditional egrep. – Grecian