Database Reference
In-Depth Information
ferent: here, we are giving a list of
username
values, and a list of
id
values, and asking
for any row that has any arbitrary pair of the given values as a primary key.
This works in the preceding case because the
id
column, as a UUID, is a globally unique
identifier for the row. We only specify the
username
column so that Cassandra knows
which partitions to go looking for the rows in; as far as identifying the row goes, the
username
is redundant information. If our data structure were such that it were possible
for rows with different partition key values to have the same clustering column value, a
query of the preceding form would be incorrect.
In this case, though, we get the results we're looking for:
