awk and equivalence classes

Question

Does gnu awk support POSIX equivalence classes?

Is it possible to match [[=a=]] using awk as it is done in grep?

$ echo ábÅ | grep [[=a=]]
ábÅ

$ echo ábÅ | grep -o [[=a=]]
á
Å

James Brown · Answer 1 · 2017-03-18T09:27:00.077

5

See here, towards the end:

Locale-specific names for a list of characters that are equal. The name is enclosed between ‘[=’ and ‘=]’. For example, the name ‘e’ might be used to represent all of “e,” “ê,” “è,” and “é.” In this case, ‘[[=e=]]’ is a regexp that matches any of ‘e’, ‘ê’, ‘é’, or ‘è’.

These features are very valuable in non-English-speaking locales.

CAUTION: The library functions that gawk uses for regular expression matching currently recognize only POSIX character classes; they do not recognize collating symbols or equivalence classes.

edited Mar 18 '17 at 09:27

answered Mar 18 '17 at 09:16

James Brown

36,089
7
43
59

1

Of course, I've found it. I cannot figure out though, whether it's possible to use it with gawk. – Eugene Barsky Mar 18 '17 at 09:20
1

Yeah, just started fighting with it myself and apparently not. – James Brown Mar 18 '17 at 09:23

score 5 · Accepted Answer · answered Mar 18 '17 at 09:17

5

Per the GAWK User's Guide, "Caution: The library functions that gawk uses for regular expression matching currently only recognize POSIX character classes; they do not recognize collating symbols or equivalence classes.".

Accordingly, you're going to have to write-out the allowed equivalents in the regex /[aáÅ]/ or whatever you're looking for.

There are locale-aware character ranges but that doesn't seem to be what you're asking about.

answered Mar 18 '17 at 09:17

Raymond Hettinger

216,523
63
388
485

1

So, it's impossible? That's a pity... :( – Eugene Barsky Mar 18 '17 at 09:20
2

It is a bummer. Until the underlying libraries get enhanced, you either have to write-out the equivalences by hand or shell-out to let another program do the work. – Raymond Hettinger Mar 18 '17 at 09:24

score 0 · Answer 3 · answered Apr 01 '22 at 05:25

You'll be surprised what gawk is willing to do these days :

 echo 'eÅêéAEè' \
                 \
 | mawk 'BEGIN { FS=RS="^$"
                   ORS=  ""
   } sub(/[\n]$/,"") +\
    gsub("[ \t]+|[\000-\b\v-\37!-\177]|"\
                 "[\200-\277]+","&\n")'  \
                                          \
 | gtee >( gpaste -s -d':' - | ecp >&2; )  \
                                            \ 
 | LC_ALL=C gawk -b -e '/[=[:lower:]=]|[=Å=]/'

e:Å:ê:é:A:E:è
```
 1   e
 2   Å
 3   ê
 4   é
 5   è
```

Even when I forced both non-multibyte "C" locale as well as using the byte mode flaw in gawk, it's willing to match it at the larger class level. However, it's unwilling to match the ASCII "A" if I only specified just the Scandinavian A-ring.

awk and equivalence classes

3 Answers3