Compressed/slow/row-oriented storage for documents.
A field needs to be marked as stored in the schema in
order to be handled in the
Internally, documents (or rather their stored fields) are serialized to a buffer.
When the buffer exceeds 16K, the buffer is compressed using
and the resulting block is written to disk.
One can then request for a specific
A skip list helps navigating to the right block,
decompresses it entirely and returns the document within it.
If the last document requested was in the same block, the reader is smart enough to avoid decompressing the block a second time, but their is no real uncompressed block cache.
A typical use case for the store is, once the search result page has been computed, returning the actual content of the 10 best document.
Most users should not access the
and should rely on either
Reads document off tantivy's