How to get just the size of the value in BerkeleyDB?

Question

Is there a way to get only the length (in bytes) of a value stored in BDB? I don't need the entire data array, only its size.

Answer 1

If you don't want to have to retrieve the entire entry and aren't using DPL, I'd say you should add a secondary index on the size of the stored byte array and make sure that your DAO properly updates this value on any save or updates. You could add a KeyCreator which creates a secondary size key in a secondary database based on the record.

What type of query are you trying to perform? Are you wanted to search for all records of a given size? Or are you wanting to know the size of a certain record before you retrieve it? I think the latter question is harder to answer.

Answer 2

I'm assuming you're using the JE version (or the Java binding of BDB) in which case, once you get the DatabaseEntry of the desired key, getSize() should give you what you want.

If you're using the C binding, check the DBT handle's size field.

Answer 3

If you store your document ids as duplicate data items, instead of as one blob data item value, then you can use DBC->count() to detect the number of matching documents without actually retrieving the long list of ids. Otherwise, the Berkeley DB API does not seem to support what you're asking for (even though you'd think it could be efficient for them to add it). I puzzled over this as well, and that was the solution I came up with for my own project.

Answer 4

for your problem, using the DB_DBT_PARTIAL flag ang asking for the begining of the record will provide you your first IDs and the DBT.size can be used to compute the total number of IDs.