Text-to-SQL systems translate natural language (NL) questions into SQL queries, enabling non-technical users to interact with structured data. While large language models (LLMs) have shown promising results on the text-to-SQL task, they often produce semantically incorrect yet syntactically valid queries, with limited insight into their reliability. We propose SQLens, an end-to-end framework for fine-grained detection and correction of semantic errors in LLM-generated SQL. SQLens integrates error signals from both the underlying database and the LLM to identify potential semantic errors within SQL clauses. It further leverages these signals to guide query correction. Empirical results on two public benchmarks show that SQLens outperforms the best LLM-based self-evaluation method by 25.78% in F1 for error detection, and improves execution accuracy of out-of-the-box text-to-SQL systems by up to 20%.

Database performance troubleshooting is a complex multi-step process that broadly involves three key stages- (a) Detection: determining what's wrong and when; (b) Root Cause Analysis (RCA): reasoning about why is the performance poor; (c) Resolution: identifying a fix. A plethora of techniques exist to address each of these problems, but they hardly work in real-world at scale. First, real-world customer workloads are noisy, non-stationary and quasi-periodic in nature rendering traditional detectors ineffective. Second, real-world production databases execute a highly diverse set of queries that skew the database statistics into long-tail distributions causing traditional RCA methods to fail. Third, these databases typically execute millions of such diverse queries every minute rendering traditional methods inefficient when deployed at scale.

Database research and development is heavily influenced by benchmarks, such as the industry-standard TPC-H and TPC-DS for analytical systems. However, these twenty-year-old benchmarks neither capture how databases are deployed nor what workloads modern cloud data warehouse systems face these days. In this paper, we summarize well-known, confirm suspected, and unearth novel discrepancies between TPC-H/DS and actual workloads using empirical data. We base our analysis on telemetrics from Amazon Redshift - one of the largest cloud data warehouse deployments. Among others, we show how write-heavy data pipelines are prominent, workloads vary over time (in both load and type), queries are repetitive, and how most properties of queries or workloads experience very long tailed distributions. We conclude that data warehouse benchmarks, just like database systems, need to become more holistic and stop focusing solely on query engine performance. Finally, we publish a dataset containing query statistics of 200 randomly selected Redshift serverless and provisioned instances (each) over a three-month period, as a basis for building more realistic benchmarks.

Modern data systems are typically both complex and general-purpose. They are complex because of the numerous internal knobs and parameters that users need to manually tune in order to achieve good performance; they are general-purpose because they are designed to handle diverse use cases, and therefore often do not achieve the best possible performance for any specific use case. A recent trend aims to tackle these pitfalls: instance-optimized systems are designed to automatically self-adjust in order to achieve the best performance for a specific use case, i.e., a dataset and query workload. Thus far, the research community has focused on creating instance-optimized database components, such as learned indexes and learned cardinality estimators, which are evaluated in isolation. However, to the best of our knowledge, there is no complete data system built with instance-optimization as a foundational design principle. In this paper, we present a progress report on SageDB, our effort towards building the first instance-optimized data system. SageDB synthesizes various instance-optimization techniques to automatically specialize for a given use case, while simultaneously exposing a simple user interface that places minimal technical burden on the user. Our prototype outperforms a commercial cloud-based analytics system by up to 3X on end-to-end query workloads and up to 250X on individual queries. SageDB is an ongoing research effort, and we highlight our lessons learned and key directions for future work.

Hashing is a fundamental operation in database management, playing a key role in the implementation of numerous core database data structures and algorithms. Traditional hash functions aim to mimic a function that maps a key to a random value, which can result in collisions, where multiple keys are mapped to the same value. There are many well-known schemes like chaining, probing, and cuckoo hashing to handle collisions. In this work, we aim to study if using learned models instead of traditional hash functions can reduce collisions and whether such a reduction translates to improved performance, particularly for indexing and joins. We show that learned models reduce collisions in some cases, which depend on how the data is distributed. To evaluate the effectiveness of learned models as hash function, we test them with bucket chaining, linear probing, and cuckoo hash tables. We find that learned models can (1) yield a 1.4x lower probe latency, and (2) reduce the non-partitioned hash join runtime with 28% over the next best baseline for certain datasets. On the other hand, if the data distribution is not suitable, we either do not see gains or see worse performance. In summary, we find that learned models can indeed outperform hash functions, but only for certain data distributions.

We present Sparse Numerical Array-Based Range Filters (SNARF), a learned range filter that efficiently supports range queries for numerical data. SNARF creates a model of the data distribution to map the keys into a bit array which is stored in a compressed form. The model along with the compressed bit array which constitutes SNARF are used to answer membership queries. We evaluate SNARF on multiple synthetic and real-world datasets as a stand-alone filter and by integrating it into RocksDB. For range queries, SNARF provides up to 50x better false positive rate than state-of-the-art range filters, such as SuRF and Rosetta, with the same space usage. We also evaluate SNARF in RocksDB as a filter replacement for filtering requests before they access on-disk data structures. For RocksDB, SNARF can improve the execution time of the system up to 10x compared to SuRF and Rosetta for certain read-only workloads.

Parametric query optimization (PQO) must address two problems: identify a relatively small number of plans to cache for a parameterized query (populateCache), and efficiently select the best cached plan to use for executing any instance of the parameterized query (getPlan). Our approach decouples these two decisions. We formulate populateCache as an optimization problem with the goal of identifying a set of plans that minimizes the optimizer estimated cost of queries in the log, and present an efficient algorithm. For getPlan, we leverage query logs to train machine learning (ML) models to choose the lowest optimizer-estimated cost plan from the cached plans. We conduct extensive experiments using complex parameterized queries from benchmarks and real workloads. Our algorithm for populateCache achieves low geometric mean sub-optimality (1.2) even for complex queries using relatively few plans, and scales well to large query logs. The mean latency of our ML model based getPlan technique (~ 210μ sec ) is between one to four orders of magnitude faster compared to prior PQO techniques. The mean sub-optimality is low (1.05), and the 95 th percentile sub-optimality (1.3) is between 1.1× and 25× lower compared to prior techniques. Finally, we present an efficient algorithm for getPlan that leverages execution time information in query logs to circumvent inaccuracies of the query optimizer's cost estimates.

Sorting is one of the most fundamental algorithms in Computer Science and a common operation in databases not just for sorting query results but also as part of joins (i.e., sort-merge-join) or indexing. In this work, we introduce a new type of distribution sort that leverages a learned model of the empirical CDF of the data. Our algorithm uses a model to efficiently get an approximation of the scaled empirical CDF for each record key and map it to the corresponding position in the output array. We then apply a deterministic sorting algorithm that works well on nearly-sorted arrays (e.g., Insertion Sort) to establish a totally sorted order. We compared this algorithm against common sorting approaches and measured its performance for up to 1 billion normally-distributed double-precision keys. The results show that our approach yields an average 3.38x performance improvement over C++ STL sort, which is an optimized Quicksort hybrid, 1.49x improvement over sequential Radix Sort, and 5.54x improvement over a C++ implementation of Timsort, which is the default sorting function for Java and Python.

We present a new dynamic partial-order reduction method for stateless model checking of concurrent programs. A common approach for exploring program behaviors relies on enumerating the traces of the program, without storing the visited states (aka stateless exploration). As the number of distinct traces grows exponentially, dynamic partial-order reduction (DPOR) techniques have been successfully used to partition the space of traces into equivalence classes ( Mazurkiewicz partitioning), with the goal of exploring only few representative traces from each class. We introduce a new equivalence on traces under sequential consistency semantics, which we call the observation equivalence. Two traces are observationally equivalent if every read event observes the same write event in both traces. While the traditional Mazurkiewicz equivalence is control-centric, our new definition is data-centric. We show that our observation equivalence is coarser than the Mazurkiewicz equivalence, and in many cases even exponentially coarser. We devise a DPOR exploration of the trace space, called data-centric DPOR, based on the observation equivalence. For acyclic architectures, our algorithm is guaranteed to explore exactly one representative trace from each observation class, while spending polynomial time per class. Hence, our algorithm is optimal wrt the observation equivalence, and in several cases explores exponentially fewer traces than any enumerative method based on the Mazurkiewicz equivalence. For cyclic architectures, we consider an equivalence between traces which is finer than the observation equivalence; but coarser than the Mazurkiewicz equivalence, and in some cases is exponentially coarser. Our data-centric DPOR algorithm remains optimal under this trace equivalence. Finally, we perform a basic experimental comparison between the existing Mazurkiewicz-based DPOR and our data-centric DPOR on a set of academic benchmarks. Our results show a significant reduction in both running time and the number of explored equivalence classes.