USA: +1(669)289-8683 
Sign in
- SQL is called a declarative language. Describe the pros and cons of a declarative language compared to an imperative language (e.g., C/C++)
- Hash can be slower than binary search. Describe the reason why.
- Assume you want to implement a parallel hash join. You can use 1000 CPU cores and all the data can fit in DRAM. Describe how you implement it. You need to consider not only search but also hash table construction. Locking should be necessary for parallel building, and an efficient locking system should be preferred.
- Describe why DSM is more efficient than NSM for aggregate operation.
- When you have N relational tables, how many plans do you have in theory?
- When making a query processing pipeline, if you adopt the Volcano method, then you can reduce memory consumption size, which is nice. Then, explain the negative aspect of this method.
- If you adopt a just-in-time compilation method, then you generate C language from SQL online, dynamically. The code generation time can be long, but this method is popular in modern databases. Explain the reason why.
- Explain false positive and false negative.
- In today’s world where software is eating the world, data platforms (DPFs) to manage data are critical. Both GAMA/BATH have their own DPFs. How would a country or an organization without DPF technology be in trouble? Please state your thoughts.
- Consider an application of eBPF.
