Which query gets much more expensive if you partition by a hash of the key instead of by key range?

Options

• ✗ Fetch the row for one key
• ✓ Scan all keys between two values, such as events between 10:00 and 11:00 Correct
• ✗ Insert a new row
• ✗ Count the rows on one shard

Why

Hashing scatters neighbouring keys across all shards, so a range scan must ask every shard. Range partitioning keeps them together, but risks hot spots when writes cluster at one end of the range.

Review this in your daily deck → All cards in Sharding