I have recently come across the term protein sequence identity, which is defined as the ratio between the number of matches between two amino acid sequences and the length of the alignment.
I was thinking of a hypothetical situation where I download the amino acid sequence of a particular protein from two different databases. The first database gives the length as X and the second database as Y (X<Y). But the identity between the two sequences is 100%. Is this possible given the way we calculate identity as described above?