My question is: are there actual music scores containing voices with initial grace notes groups of different lengths?
Torsten already explained why all the grace groups at the beginning of the piece should all have the same length. To address your question: It's not hard to make them effectively have the same length even if the respective total displayed lengths differ.
For example: The following leads to the well-known problem of misplaced clef/key changes:
\version "2.19.80"
\new PianoStaff <<
\new Staff \relative e' {
\key a \major
\grace { e16 a cis }
e4
}
\new Staff {
\clef bass
\key a \major
\grace e8
<e a, cis'>4
}
>>
But either of the following variants may be used (basically depending on how the rhythmic relationship between the different grace groups should be understood):
\version "2.19.80"
\new PianoStaff <<
\new Staff \relative e' {
\key a \major
\grace { e16 a cis }
e4
}
\new Staff {
\clef bass
\key a \major
\grace e8*3/2
<e a, cis'>4
}
>>
or
\version "2.19.80"
\new PianoStaff <<
\new Staff \relative e' {
\key a \major
\grace { e16 a cis }
e4
}
\new Staff {
\clef bass
\key a \major
\grace { s16 e8 }
<e a, cis'>4
}
>>
or even
\version "2.19.80"
\new PianoStaff <<
\new Staff \relative e' {
\key a \major
\grace { e16 a cis }
e4
}
\new Staff {
\clef bass
\key a \major
\grace { s8 e8*1/2 }
<e a, cis'>4
}
>>
and maybe even
\version "2.19.80"
\new PianoStaff <<
\new Staff \relative e' {
\key a \major
\grace { e16*2/3 a cis }
e4
}
\new Staff {
\clef bass
\key a \major
\grace { s16 e8*1/2 }
<e a, cis'>4
}
>>
I think you get my drift. :-)
Best
Lukas