Hello

At your question:

"These logs are written by different systems and I do not what what kind of encoding is this? "

How did you get such file ? It seems like double encoded.

You must ensure that every different system which is writting to such file are using the same encoding (maybe UTF8 is fine in your case)

Please tell us if you get the work done. I think that a double encode file is very dificult to decode correctly without doing manual conversions.

I would start filtering lines which contains letters not expected (add all the letters that you would expect on your file chomp; print if (/[^a-zA-Z0-9ÄÖÜäöüß,-_.\s]/); and then start to change specific words with s command.

Always is interesting to know how such file cames to the live ! I had some of them because a wrong play of substitution (wrong iconv calls)

Regards