the primary key is OrderingColumn, which isn't used apart from byTag queries.
at the same time , there is only a UNIQUE constraint on columns (PersistenceIdColumnName, SequenceNrColumnName) even though those columns are more often used in basic queries.
It should be evaluated whether making the tuple the clustered index of the table is an even better approach (https://www.sqlite.org/withoutrowid.html) depending on the use case, especially if one does not need the 'byTag' queries.
There are also some discrepancies between MsSql and Sqlite table schema, perhaps there are some optimizations that should be moved to sqlite (I mean the index on the sequence_nr column).
in SqliteQueryExecutor.cs (the Batch... as well)
the primary key is OrderingColumn, which isn't used apart from byTag queries. at the same time , there is only a UNIQUE constraint on columns (PersistenceIdColumnName, SequenceNrColumnName) even though those columns are more often used in basic queries.
It should be evaluated whether making the tuple the clustered index of the table is an even better approach (https://www.sqlite.org/withoutrowid.html) depending on the use case, especially if one does not need the 'byTag' queries.
There are also some discrepancies between MsSql and Sqlite table schema, perhaps there are some optimizations that should be moved to sqlite (I mean the index on the sequence_nr column).