Hello,
I am working on some experimental modifications to Cassandra and I was
wondering if someone can help me with an easy answer to this:
Say I have a column family in the system keyspace where I am keeping
some bookkeeping data (couple hundred thousand columns) at times.
I don't want to load all of the columns into memory before operating on
them, so I can slice by some limit at a time:
ColumnFamilyStore store =
Table.open(Table.SYSTEM_TABLE).getColumnFamilyStore(MY_CF);
if (store.isEmpty())
return null;
ByteBuffer rowKey = GreenCassandraConsistencyManager.MY_ROW;
DecoratedKey<?> epkey =
StorageService.getPartitioner().decorateKey(rowKey);
QueryFilter filter = QueryFilter.getSliceFilter(epkey, path,
startColumn, endColumn, false, limit);
ColumnFamily cf =
ColumnFamilyStore.removeDeleted(store.getColumnFamily(filter),
(int)(System.currentTimeMillis() / 1000));
My question is, can I do something similar if I have thousands of
*sub*columns. I am doing this:
endpointColumn.getColumn(columnName).getSubColumns()), which
returns a structure of all the subcolumns.
But this seems like it would load everything into memory at once. Can I
easily slice this like I slice the regular columns?
Thanks,
Bill Katsak
