Hi all,
Due to utf16/8-mismatch, I find a lot of utf16 \xnn chars in my events; this makes the json-parser kind of losing it.
So I want to get the right utf8 chars out of a dictionary json-table by doing:
f=replace(_raw,"\\\\x([0-9a-fA-F]{2})",json_extract(utfx,"{}.\1"))
The dictionary simply looks like [{"00":"utf8char-1"}, ..., {"AE":"é"},...]
But this doesn't seem to work, the event even gets nilled completely.
Something explicit like this does seem to work though: (here for instance, all utf16 \xAE chars get replaced by the "é" char:
f=replace(_raw,"\\\\x([0-9a-fA-F]{2})",json_extract(utfx,"{}.9E"))
or this:
f=replace(_raw,"\\\\x([0-9a-fA-F]{2})","\1")), which simply removes the "\x"
...so is it like the capt.groups of the regex in replace() is not evaluated if it is arg to another function io a plain string?
Tx.
