Abstract: Techniques related to binary encoding of hierarchical data objects to support efficient path navigation of the hierarchical data objects are disclosed. A hierarchical data object may include field names that are associated with field values. A method may involve generating a plurality of hash codes, each hash code corresponding to a respective field name. The method may involve generating a hash-code mapping that maps each hash code to a respective field-name identifier. The method may involve generating a field-name mapping that maps each field name to a respective field-name identifier. The method may involve generating a hierarchical tree of nodes that includes non-leaf nodes and leaf nodes. A particular non-leaf node may include a child node mapping that maps the particular non-leaf node to one or more child nodes and may include a field-name-identifier-to-child mapping that maps a respective field-name identifier to each of the one or more child nodes.

Abstract: Techniques are provided for de-normalizing semi-structured hierarchical data into a virtual table. At least a portion of semi-structured data document collection is denormalized for improving the execution of queries that involves a traversal of the semi-structured data hierarchy of the semi-structured data document collection, in an embodiment. Based on the extracted schema of the semi-structured data, a de-normalized arrangement is generated, in which the hierarchical relationship of the semi-structured data is converted into a set of columns. The denormalized arrangement is materialized by applying the de-normalized arrangement onto the semi-structured data. The materialized arrangement, the virtual table, may be stored on a persistent storage or kept in volatile memory. The virtual table may be stored in one format on the persistent storage and in another format in the volatile memory.

Abstract: Hierarchical data objects are indexed using an index referred to herein as a hierarchy-value index. A hierarchy-value index has, as index keys, tokens (tag name, a word in node string value) that are extracted from hierarchical data objects. Each token is mapped to the locations that correspond to the data for the token in hierarchical data objects. A token can represent a non-leaf node, such as an XML element or a JSON field. A location can be a region covering and subsuming child nodes. For a token that represents a non-leaf node, a location to which the token is mapped contains the location of any token corresponding to a descendant node of the non-leaf node. Thus, token containment based on the locations of tokens within a hierarchical data object may be used to determine containment relationships between nodes in a hierarchical data object.

Abstract: Techniques are provided for de-normalizing semi-structured hierarchical data into a virtual table. At least a portion of semi-structured data document collection is denormalized for improving the execution of queries that involves a traversal of the semi-structured data hierarchy of the semi-structured data document collection, in an embodiment. Based on the extracted schema of the semi-structured data, a de-normalized arrangement is generated, in which the hierarchical relationship of the semi-structured data is converted into a set of columns. The denormalized arrangement is materialized by applying the de-normalized arrangement onto the semi-structured data. The materialized arrangement, the virtual table, may be stored on a persistent storage or kept in volatile memory. The virtual table may be stored in one format on the persistent storage and in another format in the volatile memory.

Abstract: Techniques related to binary encoding of hierarchical data objects to support efficient path navigation of the hierarchical data objects are disclosed. A hierarchical data object may include field names that are associated with field values. A method may involve generating a plurality of hash codes, each hash code corresponding to a respective field name. The method may involve generating a hash-code mapping that maps each hash code to a respective field-name identifier. The method may involve generating a field-name mapping that maps each field name to a respective field-name identifier. The method may involve generating a hierarchical tree of nodes that includes non-leaf nodes and leaf nodes. A particular non-leaf node may include a child node mapping that maps the particular non-leaf node to one or more child nodes and may include a field-name-identifier-to-child mapping that maps a respective field-name identifier to each of the one or more child nodes.

Abstract: A method, device, and non-transitory computer-readable storage medium are provided for efficiently registering a relational schema. In co-compilation and data guide approaches, a subset of entities from schema descriptions are selected for physical registration, and other entities from the schema descriptions are not physically registered. In the co-compilation approach, a first schema description references a second schema description, and the subset includes a set of entities from the second schema description that are used by the first schema description. In the data guide approach, the subset includes entities that are used by a set of structured documents. In a pay-as-you-go approach, schema registration includes logically registering entities without creating relational database structures corresponding to the entities. A database server may execute database commands that reference the logically registered entities.

Abstract: Hierarchical data objects are indexed using an index referred to herein as a hierarchy-value index. A hierarchy-value index has, as index keys, tokens (tag name, a word in node string value) that are extracted from hierarchical data objects. Each token is mapped to the locations that correspond to the data for the token in hierarchical data objects. A token can represent a non-leaf node, such as an XML element or a JSON field. A location can be a region covering and subsuming child nodes. For a token that represents a non-leaf node, a location to which the token is mapped contains the location of any token corresponding to a descendant node of the non-leaf node. Thus, token containment based on the locations of tokens within a hierarchical data object may be used to determine containment relationships between nodes in a hierarchical data object.

Abstract: The approaches described herein provide an efficient way for a database server to support storage and retrieval for any of a growing number of semi-structured data formats. In one embodiment, a set of generic semi-structured data operators are provided that enable users to query, update, and validate data stored in any of a number of semi-structured data formats. In this context, a “generic” semi-structured data operator means a data operator that may be configured to operate on any number of different semi-structured data formats. For example, according to one embodiment, the same set of generic semi-structured data operators may be used to operate on data stored according to the XML, JSON, or any number of other semi-structured data formats.

Abstract: A method, device, and non-transitory computer-readable storage medium are provided for efficiently registering a relational schema. In co-compilation and data guide approaches, a subset of entities from schema descriptions are selected for physical registration, and other entities from the schema descriptions are not physically registered. In the co-compilation approach, a first schema description references a second schema description, and the subset includes a set of entities from the second schema description that are used by the first schema description. In the data guide approach, the subset includes entities that are used by a set of structured documents. In a pay-as-you-go approach, schema registration includes logically registering entities without creating relational database structures corresponding to the entities. A database server may execute database commands that reference the logically registered entities.

Abstract: A method, device, and non-transitory computer-readable storage medium are provided for efficiently registering a relational schema. In co-compilation and data guide approaches, a subset of entities from schema descriptions are selected for physical registration, and other entities from the schema descriptions are not physically registered. In the co-compilation approach, a first schema description references a second schema description, and the subset includes a set of entities from the second schema description that are used by the first schema description. In the data guide approach, the subset includes entities that are used by a set of structured documents. In a pay-as-you-go approach, schema registration includes logically registering entities without creating relational database structures corresponding to the entities. A database server may execute database commands that reference the logically registered entities.

Abstract: A method and apparatus is provided for efficiently searching and navigating XML data stored in a relational database. When storing a collection of XML documents, certain scalar elements may be shredded and stored in a relational table, whereas unstructured data may be stored as a CLOB or BLOB column. The approach includes identifying a reference address to within an XML tree index entry and storing the address in an xmltable index. The tree index entry allows for navigation in all axes. A path-based expression may be evaluated in the context of the reference address of the LOB. The result of the evaluation identifies another XML tree index entry containing a LOB locator used to retrieve the content from the document. The tree index, node index, and secondary function indexes are used together to enhance the performance of querying the XML data.

Abstract: A method, device, and non-transitory computer-readable storage medium are provided for efficiently registering a relational schema. In co-compilation and data guide approaches, a subset of entities from schema descriptions are selected for physical registration, and other entities from the schema descriptions are not physically registered. In the co-compilation approach, a first schema description references a second schema description, and the subset includes a set of entities from the second schema description that are used by the first schema description. In the data guide approach, the subset includes entities that are used by a set of structured documents. In a pay-as-you-go approach, schema registration includes logically registering entities without creating relational database structures corresponding to the entities. A database server may execute database commands that reference the logically registered entities.

Abstract: A method and apparatus is provided for efficiently searching and navigating XML data stored in a relational database. When storing a collection of XML documents, certain scalar elements may be shredded and stored in a relational table, whereas unstructured data may be stored as a CLOB or BLOB column. The approach includes identifying a reference address to within an XML tree index entry and storing the address in an xmltable index. The tree index entry allows for navigation in all axes. A path-based expression may be evaluated in the context of the reference address of the LOB. The result of the evaluation identifies another XML tree index entry containing a LOB locator used to retrieve the content from the document. The tree index, node index, and secondary function indexes are used together to enhance the performance of querying the XML data.