Using variables inside regexp constants

Faki · 03-05-2023, 04:03 PM

Can one introduce variables inside awk regexp constants ?

Code:

gsub(/^[avar]|[bvar]$/, "", str)

xode · 03-05-2023, 04:11 PM

Yes: https://stackoverflow.com/questions/...e-inside-regex But, you do need to be careful about syntax and the like.

Faki · 03-05-2023, 04:39 PM

How does it work with awk ?

MadeInGermany · 03-06-2023, 12:29 AM

Not in the / / direct RE context, but in a " " string context.

Code:

gsub(("^" avar "|" bvar "$"), "", str)

With " " you can do string concatenation and all string functions. The disadvantage is the \ must be \ escaped (i.e. doubled) when used in (casted to) a RE context:
\\ in " "
\ in / /

And of course the delimiters must be \ escaped:
backslash" in " "
backslash/ in / /
(Have to write backslash here. I mean \ )

pan64 · 03-06-2023, 12:53 AM

that is quite simple, you can use variables:

Code:

# in an awk script
regex_var="adff  fgER gqRg er" # here you can construct the regex as a normal string
gsub(regex_var, "", str)

here is a very good guide: https://learnbyexample.github.io/lea...pressions.html
look for dynamic regexp at the end.

Faki · 03-06-2023, 06:33 AM

I know how to construct regex string with variable replacements. But cannot do the same with regexp constants, it seems. Is this correct conclusion ?

pan64 · 03-06-2023, 08:45 AM

I don't understand what do you mean by that.

Faki · 03-06-2023, 01:05 PM

Meaning that a Dynamic Regex ("drgx") is superior than Constant Regex (/crgx/). Because the former allows variable substitution.