Modern database systems offer several alternative representations of the same data, both within one model and across multiple models. These choices have a strong influence on performance. But manually identifying an efficient mapping is challenging, especially in multi-model environments that combine heterogeneous operators. In this paper, we present MM-mapsearch, an automatic advisor for selecting workload-aware multi-model mappings. A simplified categorical schema defines the search space and enables uniform reasoning about structural alternatives. MM-mapsearch analyses a weighted workload, explores feasible schema variants using Monte Carlo Tree Search, and estimates their performance with a plan-structured neural predictor that generalizes across relational, document, and graph systems.

Multi-Model database systems combine the advantages of traditional and NoSQL database systems. However, the management of these systems is challenging, as users have to design an appropriate storage strategy for their data. One of the most influential factors in the storage strategy is the selection of indexes. Indexes can significantly improve query performance, but they require additional storage space and maintenance overhead. Index selection problem is well-studied in the context of single-model Database Management Systems (DBMSs), but there is a lack of research in the context of multi-model database systems. We address this problem by conducting a survey of current state-of-the-art index selection algorithms and evaluating their applicability to other DBMSs. The results reveal the strengths and weaknesses of existing algorithms and highlight the need for specialized algorithms for multi-model database systems. Moreover, we formulate open questions and suggest future research directions in this field. Our research provides a foundation for the development of efficient index selection algorithms for multi-model DBMSs.

Frequent changes in application requirements demand not only schema and query adaptation but also migration and optimization of the object-relational mapping (ORM) code. While database and query migration are well-studied, application-level translation across ORM frameworks remains largely overlooked. We present ORMorpher, a unified and extensible framework for translating and optimizing across heterogeneous ORMs. Unlike existing solutions, ORMorpher supports both structural code transformation and resource-aware framework selection under user-defined constraints. Although broadly applicable, we demonstrate its effectiveness on three widely-used .NET frameworks: Entity Framework Core, NHibernate, and Dapper over Microsoft SQL Server, enabling practical, performance-driven migration. The source code is available at: https://github.com/milan252525/orm-convertor. A demonstration video is available at: https://youtu.be/zwGGdqXtrzM.

Functional dependency (FD) discovery is fundamental to data profiling. However, existing discovery approaches can yield fake FDs that hold only coincidentally. Moreover, they often fail to identify ghost FDs that would be observable in a complete dataset but remain undetected in the available sample due to outliers. We introduce an interactive method that augments an Armstrong relation with additional tuples to eliminate fake FDs and reveal ghost FDs. Our approach utilizes artificially constructed negative examples that emulate real-world tuples to approximate dataset completeness and enhance detection accuracy. By integrating human feedback, FDepHunter provides a novel approach that refines its results iteratively, enabling more accurate identification of genuine FDs and broader applicability in real-world scenarios.

Multi-model data encompasses structurally distinct data, including relational, document, graph, key/value, columnar, etc., managed within a single system, such as a multi-model database or a data lake. Querying multi-model data requires strategies that balance unification and integration across diverse models and query languages. This paper presents DortDB, an extensible framework enabling cross-model queries combining various well-known query languages and offering intuitive flexibility, format transformation, and optimization via a unified algebra. Though a small-scale in-memory prototype is to be demonstrated, its principles can be extended to distributed systems.

This paper tackles a critical challenge in data management: adapting to evolving user requirements, a concept we term evolution management. Our approach specifically addresses the evolution management of multi-model data. By leveraging category theory, we provide an abstract representation of the combined models. We extend this categorical framework by introducing schema-modification operations of varying types and complexities. And, we propose corresponding propagation strategies to adapt queries, with a particular focus on mitigating the impact of decreasing information capacity. Through experimental verification, we demonstrate significant reductions in effort across different strategies and their effects on querying.

The rise of Big Data has exposed the limitations of relational databases in handling large datasets, driving the growth of NoSQL databases. Today, various database systems based on distinct models - or their combinations - are available, raising the question of which is best suited for a specific use case. While several papers compare subsets of these systems, they are often limited in scope. In this paper, we offer a comprehensive comparison of six systems, representing all major data models, through both static and dynamic analysis. We demonstrate their strengths and weaknesses across several realistic use cases.

Numerous challenges and open problems have appeared with the dawn of multi-model data. In most cases, single-model solutions cannot be straightforwardly extended, and new, efficient approaches must be found. In addition, since there are no standards related to combining and managing multiple models, the situation is even more complicated and confusing for users. This paper deals with the most important aspect of data management - querying. To enable the user to grasp all the popular models, we base our solution on the abstract categorical representation of multi-model data, which can be viewed as a graph. To unify the querying of multi-model data, we enable the user to query the categorical graph using a SPARQL-based model-agnostic query language called MMQL. The query is then decomposed and translated into languages of the underlying systems. The intermediate results are then combined into the final categorical result that can be expressed in any selected format. The support for cross-model redundancy enables one to create distinct query plans and choose the optimal one. We also introduce a proof-of-concept implementation of our solution called MM-quecat.

In this paper, we target the critical problem of data management - changing user requirements that need to be reflected in data structures, so-called evolution management. We introduce the first approach to evolution management of multi-model data, considering all popular models and known types of their combinations. We utilize category theory to represent all specifics of the combined models and extend it with schema-modification operations together with respective propagation strategies.

As multi-model databases become increasingly prevalent, the evolving multi-model schemas pose a significant challenge to query validity. This demo paper introduces MM-evoque, a novel tool designed to propagate schema changes across multi-model databases and synchronise not only data but also respective queries to remain valid and performant. To cover all the popular models and their combinations, MM- evoque is based on the categorical representation of multi-model data. To cover non-standard multi-model query languages of the underlying database systems, we use a conceptual multi- model query language MMQL. This demo paper introduces the functionality of MM-evoque in several types of use cases.

The European Union has recently outlined the idea of data spaces, which require complex support for various aspects of data management, such as specification, publication, storage, access, etc. In this paper, we introduce Atlas, an extensible toolset integrating techniques and approaches from two worlds - interoperable model-driven data exchange and multi-model data management. On the one hand, the toolset supports consistent model-driven authoring of data specifications based on common semantic vocabularies, including their technical artifacts such as data schemas for CSV, XML and JSON, and standards-based ontology mappings to ensure interpretability of the data as RDF. Data producers then provide their data according to the specifications, in a format with which they are the most comfortable. On the other hand, Atlas enables the users to efficiently store and query the provided data using multi-model databases. It does so by exploiting the compliance of the provided data with the data specification for minimizing the manual effort needed to store the data in the data consumer's database, regardless of the data format chosen by the data producer.

The variety feature of Big Data, represented by multi-model data, has brought a new dimension of complexity to data management. The need to process a set of distinct but interlinked models is a challenging task. In our demonstration, we present our prototype implementation MM-infer that ensures inference of a common schema of multi-model data. It supports popular data models and all three types of their mutual combinations, i.e., inter-model ref- erences, the embedding of models, and cross-model redundancy. Following the current trends, the implementation can efficiently process large amounts of data. To the best of our knowledge, ours is the first tool addressing schema inference in the world of multi-model databases.

The variety feature of Big Data, represented by multi-model data, has brought a new dimension of complexity to all aspects of data management. The need to process a set of distinct but interlinked data models is a challenging task. In this paper, we focus on the problem of inference of a schema, i.e., the description of the structure of data. While several verified approaches exist in the single-model world, their application for multi-model data is not straightforward. We introduce an approach that ensures inference of a common schema of multi-model data capturing their specifics. It can infer local integrity constraints as well as intra- and inter-model references. Following the standard features of Big Data, it can cope with overlapping models, i.e., data redundancy, and it is designed to process efficiently significant amounts of data.To the best of our knowledge, ours is the first approach addressing schema inference in the world of multi-model databases.

The support for multi-model data has become a standard for most of the existing DBMSs. However, the step from a conceptual (e.g., ER or UML) schema to a logical multi-model schema of a particular DBMS is not straightforward. In this paper, we extend our previous proposal of multi-model data representation using category theory for transformations between models. We introduce a mapping between multi-model data and the categorical representation and algorithms for mutual transformations between them. We also show how the algorithms can be implemented using the idea of wrappers with the interface published but specific internal details concealed. Finally, we discuss the applicability of the approach to various data management tasks, such as conceptual querying.

In this paper, we focus on the problem of evolution management of multi-model data. With the changing user requirements, the schema and the data need to be adapted to preserve the expected functionality of a multi-model application. We introduce a tool MM-evocat based on utilising the category theory. We show that the core of the tool, i.e., the categorical representation of multi-model data, enables us to grasp all the specifics of the individual models and their possible combinations. Its simple but powerful formal basis enables unique and robust support for evolution management.

The knowledge of a structural schema of data is a crucial aspect of most data management tasks. Unfortunately, in many real-world scenarios, the data is not accompanied by it, and schema-inference approaches need to be utilised. In this paper, we focus on a specific and complex use case of multi-model data where several often contradictory features of the combined models must be considered. Hence, single-model approaches cannot be applied straightforwardly. In addition, the data often reach the scale of Big Data, and thus a scalable solution is inevitable. In our approach, we reflect all these challenges. In addition, we can also infer local integrity constraints as well as intra- and inter-model references. Last but not least, we can cope with cross-model data redundancy. Using a set of experiments, we prove the advantages of the proposed approach and we compare it with related work.

Multi-model data is organised in various mutually interlinked formats and models, often with contradictory features. In addition, its structure may change over time, and its size can grow to the extremes of Big Data. In terms of research and practical processing, this creates one of the most complex challenges of effective data management. As it is not humanly possible to handle such a complex task manually, in this vision paper, we focus on the area of automatic management of dynamic multi-model Big Data. We envision a framework capable of accepting different levels of user input and different types of data, queries, changes, and propagation strategies and ensuring the preservation of adequate and efficient data access.

The variety feature of Big Data, represented by multi-model data, has brought a new dimension of complexity to data management. The need to process a set of distinct but interlinked models is a challenging task. In our demonstration, we present our prototype implementation MM-infer that ensures inference of a common schema of multi-model data. It supports popular data models and all three types of their mutual combinations, i.e., inter-model ref- erences, the embedding of models, and cross-model redundancy. Following the current trends, the implementation can efficiently process large amounts of data. To the best of our knowledge, ours is the first tool addressing schema inference in the world of multi-model databases.

NoSQL databases are becoming increasingly more popular due to their undeniable advantages in the context of storing and processing Big Data, mainly horizontal scalability and minimal requirement to define a schema upfront. In the absence of the explicit schema, however, an implicit schema inherent to the stored data still exists and it needs to be reverse engineered from the data. Once inferred, it is of a great value to the stakeholders and database maintainers. Nevertheless, the problem of schema inference is non-trivial and is still the subject of ongoing research. In this paper we provide a comparative analysis of five recent proposals of schema inference approaches targeting the JSON format. We provide both static and dynamic comparison of the approaches. In the former case we compare various features. In the latter case we involve both functional and performance analysis. Finally, we discuss remaining challenges and open problems.

The support for multi-model data has become a standard for most of the existing DBMSs. The tools for database design are general enough to cover multiple models, too. However, the step from a conceptual (e.g., ER or UML) schema to a logical multi-model schema of a DBMS (or their combination) is not straightforward. In this paper, we show how category theory can be used for representation of multi-model data and schema and how the mutual mapping between the categorical representation and logical models of particular DBMSs can be designed. For this purpose we define the notion of an access path which enables to specify the mapping for all the currently popular models. To demonstrate advantages of the proposal we introduce framework MM-cat which applies the proposed approaches on MongoDB and PostgreSQL.

With the advent of multi-model database management systems, the boundaries of many approaches to data processing were pushed. The aspect of multi-model data introduces a new dimension of complexity and new challenges not seen in single-model systems. We have to address issues arising from the combination of interconnected and often contradictory logical models, such as, e.g., order-preserving/-ignorant, aggregate-oriented/-ignorant, schema-full/-less/-mixed approaches, intra- and inter-model references, intra- and inter-model integrity constraints, or full and partial intra- and inter-model data redundancy. Hence, a number of mature and verified approaches for various data management tasks commonly used for single-model DBMSs cannot be directly applied to multi-model DBMSs. This thesis aims to propose a new family of unified approaches for both conceptual and logical multi-model modelling and data management. We first analyse the state-of-the-art of related areas. Then we propose abstract data structures to represent multi-model schema and data. These structures are then utilised in the design of approaches for unified schema inference, data migration, schema evolution, and correct backward propagation of changes to the data. All the proposed approaches are implemented and experimentally verified.

The support for multi-model data has become a standard for most of the existing DBMSs. However, the step from a conceptual (e.g., ER or UML) schema to a logical multi-model schema of a particular database, as well as its further correct maintenance, is not simple and straightforward.In this paper, we introduce MM-cat, a modular and extensible framework that enables to perform all the necessary steps to design a multi-model schema. An initial ER schema is transformed to a unified categorical representation which can then be easily mapped to any combination of models in a particular DBMS, or their set. In addition, MM-cat also provides a categorical representation of multi-model data instances to be processed uniformly. Both the schema and instance categories serve as the core for further management of multi-model data. The whole process of transformation to/from the categorical representation is demonstrated using MongoDB and PostgreSQL.

Following the current trend, most of the well-known database systems, being relational, NoSQL, or NewSQL, denote themselves as multi-model. This industry-driven approach, however, lacks plenty of important features of the traditional DBMSs. The primary problem is a design of an optimal multi-model schema and its sufficiently general and efficient representation. In this paper, we provide an overview and discussion of the promising approaches that could potentially be capable of solving these issues, along with a summary of the remaining open problems.

In this vision paper, we introduce an idea of a framework that would enable us to model, represent, and manage multi-model data in a unified and abstract way. Its core idea exploits constructs provided by category theory, which is sufficiently general but still simple enough to cover any of the logical data models used in contemporary databases. Focusing on promising features and taking into account mature and verified principles, we overview the key parts of the framework and outline open questions and research directions that need to be further investigated. The ultimate objective is to pursue the idea of a self-tuning system that would permit us to collapse the traditionally understood conceptual and logical layers into just a single model allowing for unified handling of schemas, data instances, as well as queries.

As most of the DBMSs have become multi-model, there have occurred plenty of related issues. One of them is a design of a multi-model application, where the step from the conceptual layer to a set of distinct interlinked logical models is not straightforward. We propose an approach based on category theory, which provides a unified view of the data and a solid mathematical basis for their management. We propose a schema and instance categories covering popular models and we show how an ER model can be transformed to such a categorical layer. We also outline the whole framework based on the categorical model.

In recent years, many so-called multi-model database management systems have emerged, mainly as extensions of the existing single-model systems, regardless they used to be relational or NoSQL. These new database systems make new demands on their users. From the point of view of the conceptual and logical representation, the so far widely used approaches, especially ER and UML, prove not to be sufficient enough in many aspects due to the specific properties of multi-model data. In addition, it is also difficult to query data that is represented in various and often overlapping data models at the logical level.

Since the traditional relational database systems are not capable of following the contemporary requirements on Big Data processing, a family of NoSQL databases emerged. It is not an exception for such systems not to require an explicit schema for the data they store. Nevertheless, application developers must maintain at least the so-called implicit schema. In certain situations, however, the presence of an explicit schema is still necessary, and so it makes sense to propose methods capable of schema inference just from the structure of the available data. In the context of document NoSQL databases, namely those assuming the JSON data format, we focus on several representatives of the existing inference approaches and provide their thorough comparison. Although they are often based on similar principles, their features, support for the detection of references, union types, or required and optional properties differ greatly. We believe that without adequately tackling their disadvantages we identified, uniform schema inference and modeling of the multi-model data simply cannot be pursued straightforwardly.

Distributed systems such as Apache Hadoop are optimized primarily for read operations. We generate a large amount of data that we need to efficiently archive and process, and we therefore require a steady performance of write operations that does not drop with the amount of data written. We can choose from a range of distributed systems that can be deployed on commodity hardware. These systems offer scalability not only for system load, but also for hardware resources and storage capabilities. Of course, all-time availability and resilience to network communications or parts of nodes are commonplace. We will focus on describing these distributed systems and comparing their advantages and disadvantages. We will design the solution and link the selected distributed storage to the default archive of experimental data. We gain a decentralized system that is scalable, available, fail-safe, and is not only suitable for archiving small volume experimental data but also for archiving large image and other data.