For the design principles of HBase rowkey, which of the following descriptions is correct? (multiple choice)

Question

Options

• AThe attribute value with high access weight is placed in front of RowKey
• BThe attribute value with high access weight is placed in the second half of Rowkey
• CThe attribute value with good dispersion is placed in the first half of Rowkey
• DAttribute values with good dispersion are placed in the second half of Rowkey

• A
  82% (23)
• B
  4% (1)
• D
  14% (4)

Explanation

HBase stores rows in lexicographic order and partitions data into regions based on rowkey prefixes. Placing high-access-weight attributes first (A) enables efficient prefix scans - HBase can quickly locate the target region and avoid full-table scans when the most-queried field anchors the key. Placing high-dispersion attributes first (C) prevents "hotspotting," where sequential or low-cardinality prefixes funnel all reads/writes to a single region server, creating a bottleneck.

B is wrong because burying high-access attributes in the second half forces HBase to scan more data before it can filter - you lose the benefit of prefix-based lookups. D is wrong because dispersion in the second half doesn't help distribute load across region servers; region splitting is governed by the key's prefix, so low-dispersion prefixes still cluster traffic on the same node regardless of what follows.

Memory tip: Think "Front-Loaded HBase" - put your Frequent (high-weight) and Fan-out (high-dispersion) attributes First. Both benefits (fast lookup and balanced load) come from controlling the prefix.

No community discussion yet for this question.