not a satisfying solution)

recording for build_ref_db and ecotag

columns

different views

threshold requires that they are

in bits not bytes and using view API instead of column API

properly set to 0 if the value was NA (ignore previous commit with the
same message)

double instead of int

properly set to 0 if the value was NA

for the taxonomic assignment of sequences

metadata for the taxonomic assignment of sequences

with keeping all AVLs mapped seems the most efficient. Now 1 million
sequences more or less constantly takes 1 minute.

when adding new values, fixed a bug when calculating if an AVL data file
has reached the maximum size, fixed a casting bug, and added a boolean
so read-only AVLs files are not truncated

closing and reopening indexers so that they are mapped properly, and
fixed memory leak when reading sequences

a Cython compilation bug

bugs potentially involving name conflicts

with the new versions

to use the default DMS

rest as a character string

typed views, a function to delete a view and fixed view history bug

resulting in its deletion when reopening the DMS.

of 1 (generating memory bugs when using the counter)

OBI1)