paul-rogers commented on a change in pull request #1843: DRILL-7350: Move RowSet related classes from test folder URL: https://github.com/apache/drill/pull/1843#discussion_r314399600
########## File path: exec/java-exec/src/main/java/org/apache/drill/exec/physical/rowSet/package-info.java ########## @@ -32,6 +32,22 @@ * batch" or "record batch." (But, in Drill, the term "record batch" also * usually means an operator on that set of records. Here, a row set is * just the rows &nash; separate from operations on that data.</dd> + * <dt>SingleRowSet (abstract)</dt> + * <dd>Represents a row set that contains a single record batch (the typical + * case.</dd> + * <dt>DirectRowSet</dt> + * <dd>A read-only single row set without a selection vector.</dd> + * <dt>IndirectRowSet</dt> + * <dd>A read-only, single row set with an SV2. Note that the SV2 itself is + * writable (such as for sorting.)</dd> + * <dt>ExtendableRowSet</dt> + * <dd>A write-only, single row set used to create a new row set. Because of + * the way Drill sets row counts, an extendable row set cannot be read; instead + * at the completion of the write the extendable row set becomes a direct or + * indirect row set.</dd> + * <dt>HyperRowSet</dt> + * <dd>A read-only row set made up of a collection of record batches, indexed via an + * SV4. As with the SV2, the SV4 itself is writable.</dt> Review comment: The SV is not writable within a RowSet. Changing the SV will mess up indexing. The way to use that is to have a DirectRowSet on the original batch. Create the SV. Then, use the DirectRowSet and the SV to create an indirect row set. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
