Release Notes

This page provides the release notes for Ultipa Graph Database since 2022. These notes detail major features, bug fixes, performance improvements, and other changes introduced in each version.

rc.5.1.1 (2025-05-20)

• Introduced new GQL DDL syntax, including GRAPH TYPE, HDC, PROJECTION, COMPACT, TRUNCATE, etc.
• Added backup and restore functionality supported via the ultipa_backup_client tool.
• Added built-in date functions.
• Optimized query executor to improve performance and resource efficiency.
• Resolved known bugs

rc.5.0.1 (2025-04-03)

• Introduced vector indexing and vector search.
• Added GQL DDL syntax support for managing user, role, privilege, index, full-text index, constraint, job, and role.
• Introduced database backup features: db.backup.create() and db.backup.show().
• Added the db_backup_path configuration option to the name server and meta server.
• Enhanced support for certain features when the license is expired or encounters errors.
• Updated log naming conventions to help administrators easily identify the latest logs.

beta.5.0.64 (2025-02-25)

• Introduced new HDC algorithms: Katz Centrality and TextRank.
• Introduced new distributed algorithm: SSSP (Single Source Shortest Path).
• Added new parameters to the HDC algorithm Betweenness Centrality: edge_schema_property and impl_type.
• Added a new parameter node_label_property to the distributed algorithm LPA (Label Propagation Algorithm).
• Prevented the deletion of properties that are indexed.
• Refined configuration options for shard server, name server, and meta server.
• Improved overall system performance.

beta.5.0.55 (2025-01-14)

• Added GQL DDL syntax support for managing graph, schema, and property.
• Updated distributed algorithms: Triangle Counting and K-core.
• Fixed occasional UQL blocking issues involving multiple replicas.
• Resolved intermittent UQL blocking caused by the load balancer.
• Major improvements to access control mechanisms.
• Renamed project to projection.
• Updated configuration options for the name server and HDC server.
• Added support for dynamic synchronization in HDC graphs.
• Added support for property encryption across all value types, including AES128, AES256, RSA, and ECC.
• Introduced EDGE KEY and NOT NULL constraints for improved schema definition and data integrity.
• Process lists are now stored on the name server to reduce network overhead.

beta.5.0.33 (2024-11-12)

• UQL node deletion behavior is now consistent with GQL: nodes with adjacent edges cannot be deleted by default unless force deletion is explicitly enabled. - The decimal property value type is nowed supported.
• Optimized the management for HDC graphs.
• Simplified the syntax for running algorithms in both UQL and GQL.
• The cache.xxx.turnOn, cache.xxx.turnOff, and cache.xxx.clear commands now read the corresponding xxx_cache_size from the configuration file and reconfigure it dynamically for hot updates. The updated cache_size is also returned in cache.xxx.status.
• Modified some configuration options for the name server and shard server.
• Schema deletions are now executed as jobs.

beta.5.0.7 (2024-10-11)

• Introduced more predicates, functions, expressions, and operators to GQL.
• Introduced more statements to GQL.
• Introduced advanced configurations for CPU cores. You can now flexibly control the maximum and minimum CPU cores available based on different time ranges, as well as view the total count of CPU cores.
• Fixed known bugs.

beta.5.0.5 (2024-08-30)

• GQL is now supported.
• Introduced MATCH, CREATE and MERGE clauses to UQL.
• Introduced the Schema Overview algorithm.

v4.5.3 (2024-07-25)

• Fixed issues where the step is set to 0 in template queries.
• Refactored the SET clause.
• Fixed known bugs.

beta.5.0.4 (2024-07-04)

• Introduced three new graph algorithms: Text Rank, K1 Coloring and Conductance.
• The system property _uuid can no longer be assigned by users.
• The start and end nodes of edges cannot be altered any more.
• Introduced composite property index.

beta.5.0.3 (2024-06-26)

• Introduced the synchronization mechanism between meta server and shard.
• A UQL query now can only be concluded with the RETURN, UNION or LIMIT clause.

beta.5.0.2 (2024-06-06)

• _id and schema._id now support LTE.
• Introduced composite property index.
• The usage of @* is now disabled in index creation.
• Introduced nodetach(), delete().nodes().nodetach() will not delete edges that attached to the deleted nodes.
• Fixed crash issue caused by the deletion of graphsets.
• limit() in template queries now applies to each single start node.

beta.5.0.1 (2024-05-23)

• Fixed the crash issue occurred because some schemas did not support float data types.
• The LTE and turnOn operations have been modified to run as task.
• Introduced zip operator.
• Introduced speed cache.
• Refactored caching functionalities. The logic for adding, deleting, and modifying graphsets has been changed.

v4.4.44 (2024-05-13)

• Resolved issues regarding LTE.
• Fixed known bugs.

v4.4.41 (2024-04-23)

• Optimized memory usage of the graph topological structure.
• Enhanced the efficiency of template aggregation.
• Resolved the startup consistency issue.
• Fixed known bugs.

beta.5.0.0 (2024-04-18)

• The initial beta release of Ultipa Graph v5! Introduced shard, meta server, and hdc server.
• Task is updated to Job.

v4.4.34 (2024-03-28)

• Introduced the edge_schema_property parameter to the Closeness Centrality and Harmonic Centrality algorithms.
• Fixed known bugs.

v4.4.31 (2024-03-25)

• Resolved the issue of performing a Cartesian product when returning aggregate functions along with heterologous data.
• Optimized the filtering of the decimal type properties.
• Fixed known bugs.

v4.4.27 (2024-02-29)

• Resolved the parsing error occurred when the schema name contains the character ".".
• Fixed known bugs.

v4.4.23 (2024-02-19)

• db.backup.create() has been adjusted to be as a system privilege.
• Resolved the bug occurred when index is created for the property used in CASE.
• Fixed known bugs.

v4.4.21 (2024-01-20)

• The UNCOLLECT clause now supports multiple expressions.
• Fixed known bugs.

v4.4.20 (2023-12-27)

• Introduced SET clause, e.g., find().nodes(1) as n SET n.value = 10, n.age = 12 return n{*}.
• Fixed known bugs.

v4.4.19 (2023-12-14)

• Introduced percentile_cont and percentile_disc functions.
• Optimized toGraph function to support lists containing PATH or GRAPH type elements and to allow multiple parameters without restrictions on their order.

v4.4.15 (2023-11-21)

• Optimized the efficiency of template aggregation.
• Introduced GRAPH data structure.
• Introduced toGraph function.

v4.3.94 (2023-11-13)

• Introduced blob type to properties.
• The expireDate field in the stats() return now supports displaying "Long Term."
• Introduced new operators KhopTemplateCount, KhopTemplateGroupCount, TemplateCount, and OptionalTemplateCount. They are visible when using the EXPLAIN prefix.
• Introduced new type conversion functions: toSet(), toDouble() and toDecimal().
• Added extra in the returns of the show().schema() clause to show the precision and scale information of decimal properties.
• Added size in the returns of the show().index() clause to show the size of the index in bytes.
• Introduced the TRY prefix.
• New supported syntax: n(<filter?> as nodes) as paths return nodes{*}, paths{*}
• Updated the incremental backup functionalities:
  • Create backup: db.backup.create("<backup_name>")
  • Show backups: db.backup.show("<backup_name?>")
  • Restore backup: db.backup.restore("<backup_name>", <backup_id?>)
• Deleted rpc Backup (BackupRequest) returns (BackupReply) interface in proto.
• Introduced graph privileges CREAT_BACKUP, RESTORE_BACKUP and SHOW_BACKUP.

v4.3.80 (2023-10-30)

• Introduced incremental backup functionalities:
  • Create backup: db.backup.create("<backup_path>")
  • Show backups: db.backup.show("<backup_path>")
  • Restore backup: db.backup.restore("<backup_path>", backup_id?)
• Introduced set type to properties.

v4.3.77 (2023-10-19)

• Introduced decimal type to properties.
• Added extra in the returns of the show().property() clause to show the precision and scale information of decimal properties.
• Introduced new string functions: trim(), ltrim(), rtrim(), left(), right(), substring(), reverse() and replace().
• Introduced string concatenation using the operator +, e.g., return 'a'+'b'.

v4.3.74 (2023-09-25)

• Introduced JSON_decode and JSON_merge functions.
• Fixed known bugs.

v4.3.71 (2023-09-18)

• Introduced property encryption. E.g., create().node_property(@default, "name", string).encrypt(), encrypt() encrypts the node property @default.name with AES128. Only support for string and text types.
• Fixed issues related to trigger.

v4.3.65 (2023-08-31)

• Introduced property privileges.
• The show().property() clause now indicates whether read and write operations are permitted for each property.

v4.3.60 (2023-08-18)

• Introduced the trigger functionality.
• Introduced new features to the point type, including using . to extract the coordinate values (e.g., nodes.pointType.x, nodes.pointType.y), and new function pointInPolygon.

v4.3.58 (2023-07-28)

• Enhanced rules for template queries.
• The LPA and HANP algorithms now support multiple properties.
• Fixed some bugs in algorithms.

v4.3.56 (2023-07-19)

• Addressed some abnormal crash problems.
• Improved EXTA interface performance.
• Resolved EXTA installation issue in ARM architecture.

v4.2.66 (2023-06-27)

• Updates were made to the monitor.
• Fixed known bugs.

v4.3.51 (2023-06-13)

• Resolved the memory leak issue in the algorithms.
• Fixed known bugs.

v4.3.49 (2023-05-29)

• Aggregation functions now ignore null values.
• Added ListData tag in proto.
• Introduced new spatial function point.
• Server.config configuration parameters private_addr and public_addr now support domain names.
• Fixed the issue of memory not being released under ARM architecture by changing the memory allocator to jemalloc and limiting the huge page size.

v4.2.65 (2023-05-11)

• Resolved memory leak in "ultipa.lic" certificate timing monitoring.
• Fixed out-of-memory (OOM) bug in K-hop queries.
• Fixed crash issue in the Eigenvector Centrality algorithm.

v4.3.31 (2023-04-25)

• Fixed the bug causing random results in list filtering.
• Optimized aggregate functions.
• Fixed known bugs.

v4.3.26 (2023-04-11)

• Renamed distinct function to dedup, DISTINCT is now an operator keyword.
• Refactored the UNION, WITH and RETURN clauses.
• Resolved parsing issues related to special characters.

v4.3.22 (2023-03-31)

• Introduced is_null field in AttrListData.
• Modified the proto.
• Fixed issues related to inserting and updating properties of the point type.

v4.3.12 (2023-03-15)

• Introduced the point type to properties.
• Introduced listContains and distance functions.
• The OPTIONAL prefix now supports more clauses, including find().nodes(), find().edges() and k-hop().
• Resolved inconsistency issues between memory and disk when inserting data with a mix of _uuid and _id specifications.
• Fixed bugs related to the list type.
• Updated EXTA functionalities.
• Introduced is_null tag in proto.
• Introduced the Backup interface in proto for creating backups of the entire database to a specified directory.
• Added new startup parameter -restore in ultipa-server to restore data from backup file directory.

v4.2.59 (2023-02-23)

• Resolved crashes related to lists.
• Resolved bugs related to null values in aggregation functions.
• Merged HTAP log and server log into a single file.

v4.2.53 (2023-02-06)

• Optimized metadata structure to save memory usage.
• Improved storage and performance of modification functions.
• Introduced EXTA functionality that allows custom algorithm plugins.
• Added the text type to properties without text-length limit.
• Introduced Server.docker_mem_usage_path and Server.memory_max_limit in config for memory limit control in the Cloud version.
• Timestamp data type now allows SDK to set timezone.
• Revised naming rules for schema, property, and alias to support special characters and Chinese.
• Insert, delete, update, and some query clauses now supports the limit() method and OPTIONAL prefix.
• Introduced new functions: dateFormat(), toString(), range(), ifnull(), reduce().
• Introduced the DEBUG prefix for performance statistics with better granularity.
• Modified the PROFILE prefix functionality to display each clause's time cost.
• Enhanced the display content of EXPLAIN prefix.
• WHERE clause now supports aggregate functions.
• Modified mathematical operators and functions to handle null values.
• Fixed other bugs.

v4.3.2 (2023-01-12)

• Introduced the list type to properties.
• Introduced sub types to properties.
• Introduced null values to properties.
• Introduced IS NULL and IS NOT NULL operators.

v4.2.40 (2023-01-05)

• Refactored EXPLAIN prefix functionalities.
• Renamed functions to use CamelCase naming convention, including changing date_add to dateAdd, date_diff to dateDiff, date_format to dateFormat, day_of_week to dayOfWeek, array_union to listUnion, starts_with to startsWith and ends_with to endsWith.
• Resolved bugs in the between and listUnion functions.
• Introduced function ifNull.
• Resolved issues with empty edges being included in returned paths.
• Added an error message when truncate().graph() is applied to an incorrect graphset.

beta.4.2.35 (2022-12-08)

• Optimized the Louvain algorithm.
• Introduced the Louvain Serial algorithm.
• Optimized the algorithm efficiency when _id is used as label in LPA.
• Improved the error message when the specified property to be LTE-ed does not exist.
• Fixed known bugs.

beta.4.1.53 (2022-11-16)

• Resolved the crash issue in the K-Core and Subgraph algorithms.
• Resolved the return issue of the alter() clause where the specified property does not exist.
• Fixed known bugs.

beta.4.1.45 (2022-10-24)

• Optimized the file writeback feature in algorithms.
• Resolved the issue where the EXPALIN prefix does not print aggregate functions.
• Resolved the issue where the tasks still in the writing status are stopped by the clear().task() clause.
• Fixed known bugs.

beta.4.1.31 (2022-09-20)

• Optimized the Degree Centrality algorithm.
• Resolving the slow speed issue in the file writeback for the Similarity algorithm.
• Fixed known bugs.

beta.4.1.27 (2022-09-02)

• Optimized the PATH structure.
• Refactored the find() clause.
• Fixed known bugs.

beta.4.1.21 (2022-08-10)

• Optimized the NODE and EDGE structure.
• Resolved the issue where _id is empty in the return of UNION and UNION ALL clauses.
• Resolved the property writeback failure for the CELF algorithm.

beta.4.1.8 (2022-07-19)

• Introduced the Dijkstra's Single-Source Shortest Path, Delta-Stepping Single-Source Shortest Path, and SPFA algorithms.
• Resolved the crash issue for the LINE algorithm.
• Fixed known bugs.

beta.4.0.233 (2022-07-12)

• Updated the no_circle() parameter of the ab() clause.
• Introduced the Harmonic Centrality, Eigenvector Centrality and HITS algorithms.
• Fixed known bugs.

beta.4.0.232 (2022-06-30)

• Resolved the crash issue in LPA.
• Optimized the array_union function.
• Fixed an issue that occurred when attempting to uninstall an algorithm that was still running.
• Fixed known bugs.

beta.4.0.220 (2022-06-07)

• Resolved the issue where the count reported by show().schema() was incorrect after deleting edges.
• Resolved the crash issue in the GraphSAGE Train algorithm.

beta.4.0.202 (2022-05-06)

• Resolved the wrong return issue when the first augment of the coalesce function is null.
• Resolved the issue where PATH is used in the GROUP BY clause.
• Fixed known bugs.

beta.4.0.187 (2022-04-15)

• Resolved the issue when the ORDER BY clause is used with the GROUP BY clause.
• Resolved the crash issue in the Random Walk algorithm.
• Fixed known bugs.

beta.4.0.164 (2022-03-14)

• Optimized the sampling feature for the Betweenness Centrality algorithm.
• Optimized the limit() parameter.
• Fixed known bugs.

beta.4.0.152 (2022-02-09)

• Optimized multithreading concurrency in the cluster.
• Optimized the precision of the data returned by algorithms.
• Fixed known bugs.

beta.4.0.145 (2022-01-11)

• Resolved the crash issue in the Triangle Counting algorithm.
• Optimized the stats return of the Louvain and LPA algorithms.
• Fixed known bugs.