Practice Test 3 | Google Cloud Certified Professional Data Engineer | Dumps | Mock Test
You have an on-premise relational database which you want to migrate to Google Cloud. You choose Cloud Spanner for importing the database to. You want users to be fully benefited from Cloud Spanner by ensuring the best performance. Which of the following should be taken into consideration while migrating your tables to Cloud Spanner?
A. Use version 4 UUID as primary keys for your tables. Keep the original primary keys for legacy queries.
B. Make sure your tables primary keys are monotonically increased.
C. Use UNIX timestamp as a primary key for your tables. Keep the original primary keys for legacy queries.
D. Use the combination of a timestamp and a primary key (pk) as #timestamp-pk.
Correct Answer: A
Cloud Spanner uses compute nodes to read and write data. The data of tables is stored lexicographically by primary key. Data is distributed among multiple storage “splits.”
This is the reason why choosing the right primary key for Cloud Spanner is important for performance. If the primary key is monotonic, it leads to storing table data to one storage split, which in return leads to compute nodes to hit the same storage split for reading & writing. A good primary key is a key which helps in distributing data evenly among different storage splits.
Option A suggests using version 4 UUID as a primary key. A version 4 UUID or a universally unique identifier is a 128-bit number used to identify information in computer systems. An example of a UUID is “81c96908-6a8f-46b2-bc16-3ee4c5376182” which consists of 32 hex characters.
UUIDs generate widely unique and diverse keys which allow potential primary keys, so it can be a good choice to consider to generate primary keys for records in Cloud Spanner.
Option B is incorrect: As mentioned, monotonic primary keys will lead to performance issues while reading & writing data to only one storage split.
Option C is incorrect: UNIX timestamps are monotonic since only right-most digits are changing while left-most digits will be the same in almost all cases, which leads to a performance issue.
Option D is incorrect: Combining timestamp with primary key as #timestamp-pk will lead to having PK combination with the left-most characters being the same for all primary keys.
Source(s):
Choosing the Right Primary Keys (TIL about Cloud Spanner):
https://www.youtube.com/watch?v=FFTHQt_KFNM
Universally Unique Identifier (UUID):
Comments are closed, but trackbacks and pingbacks are open.