Mini-LSM Week 1 Day3

Week1 Day3 的内容，实现 SST block 编解码和 iterator

DuckDB: an Embeddable Analytical Database

读一下 DuckDB 的论文，对 OLAP 没什么了解，但论文很短也可以看看。

CatSQL: Towards Real World Natural Language to SQL Applications

NL2SQL, text to SQL 是很有趣的方向。有 DL 方法也有现在的 LLM 微调。

The Case for a Learned Sorting Algorithm

除了 Query Optimization, Index, Tunning, ML 还可以用在 Database 其他方面，比如排序？

An Inquiry into Machine Learning-based Automatic Configuration Tuning Services on Real-World Database Management Systems

CMU 的对于 DBMS 自动调优的论文，采用了 ML 机器学习方法，是 Ottertune 的论文。

MB2: Decomposed Behavior Modeling for Self-Driving Database Management Systems

self-driving database management systems

Mini-LSM Week 1 Day2

Week1 Day2 的内容，实现 Merge Iterator

Mini-LSM Week 1 Day1

记录下 LSM 的学习过程，感谢迟先生的教程 https://skyzh.github.io/mini-lsm/

Bao: Making Learned Query Optimization Practical

MLDB + query optimization

Zero-Shot Cost Models for Out-of-the-box Learned Cost Prediction [VLDB 2022]

Abstract

本文介绍了 zero-shot cost model，该模型可以使学习的成本估算能够 generalizes to unseen databases。与最 state-of-the-art 的工作负载驱动的方法相反，这些方法不必在每个新数据基础上执行大量 training queries ，zero-shot cost models thus allow to instantiate a learned cost model out-of-the-box without expensive training data collection。为了 zero-shot cost models，本文提出 a new learning paradigm based on pre-trained cost models。As core contributions to support the transfer of such a pre-trained cost model to unseen databases, we introduce a new model architecture and representation technique for encoding query workloads as input to those models. As we will show in our evaluation, zero-shot cost estimation can provide more accurate cost estimates than state-of-the-art models for a wide range of (real-world) databases without requiring any query executions on unseen databases. Furthermore, we show that zero-shot cost models can be used in a few-shot mode that further improves their quality by retraining them just with a small number of additional training queries on the unseen database.