GraphDB Free 7.0
Table of contents
- General
- Quick start guide
- Installation
- Administration
- Usage
- Tools
- References
- Release notes
- FAQ
- Support
Release notes¶
GraphDB release notes provide information about the features and improvements in each release as well as various bugfixes. GraphDB’s versioning scheme is based on semantic versioning. The full version is composed of three components:
major
.minor
.patch
e.g., 7.0.1
where the major version is 7
, the minor version is 0
and the patch version is 1
.
Note
Releases with the same major and minor versions do not contain any new features. Releases with different patch versions contain fixes for bugs discovered since the previous minor. New or significantly changed features are released with a higher major or minor version.
GraphDB 7 includes two components with their own version numbers:
- Sesame
- GraphDB Connectors
Their individual versions use the same semantic versioning scheme as the whole product and their values are provided only as a reference.
GraphDB 7.0.3¶
Released: 16 June, 2016
Component versions¶
Sesame | Connectors |
2.8.11 | 5.0.1 |
Highlights¶
- Bugfixes in the GraphDB engine.
GraphDB Engine¶
Bug fixes¶
- OWLIM-3038 A rule line ending with literal can not have a [Cut] or [Constraint ...] modifier. If some line in a rule definition ends with literal and a modifier the rule compiler shows error. Example: c <u> “true”^^xsd:boolean [Cut]
- OWLIM-3040 Suboptimal premice order when generating rule variants for a ruleset. Some consistency rules are evaluated very slowly due to inefficiently generated code.
- OWLIM-3041 If there is an explicit statement within a context and same triple has been inferred to a system context the wildcard iterator skips it.
GraphDB Workbench¶
No changes.
GraphDB Connectors¶
No changes.
GraphDB 7.0.2¶
Released: 7 June, 2016
Component versions¶
Sesame | Connectors |
2.8.11 | 5.0.1 |
Highlights¶
- Bugfixes and improvements in the GraphDB engine;
- Bugfixes and improvements in the LoadRDF;
- Bugfixes and improvements in the GraphDB Workbench.
Note
Due to the active development and some recently discovered issues in the LoadRDF tool, some functionalities have been changed and have not been tested thoroughly. Therefore, for this release only, we recommended you to avoid ‘disable-sameAs’ option and using the ‘parallel’ mode of LoadRDF, in order to achieve consistent results.
GraphDB Engine¶
Improvements¶
- OWLIM-2872 Improve query optimisation when geo-spatial is used.
- OWLIM-2801 Extended SPARQL regression test framework to also measure execution time and full results response.
- OWLIM-1919 LoadRDF - improved behavior with bad data. LoadRDF has new options for allowing partial load of a file with a corrupt line and for interrupting the load if the dataset contains a corrupt file.
- OWLIM-3035 LoadRDF - Default repository storage folder changed from ‘owlim-storage’ to ‘storage’. Now the location of the folder is consistent across all tools and the engine.
Bug fixes¶
- OWLIM-2839 Queries with
ORDER BY
can not be interrupted. - OWLIM-2986
BIND within OPTIONAL
returns incorrectly set previous value.BIND within OPTIONAL
fails to update its value and returns incorrect results. - OWLIM-2952 The master doesn’t take into account that workers are busy with serialising results in the iterators. Most of the queries hit the first worker, because the master considers that a worker, which serialises the results of a query is in idle mode.
- OWLIM-2937 Consistency rules with multiple consequents do not trigger. GraphDB ignores consistency rules, which ensure that if a resource has a sub-property ‘p’ of ‘a:hasValue’, the resource’s class must have an OWL cardinality for ‘p’.
- OWLIM-2943 LoadRDF freezes on parsing errors in parallel mode.
- OWLIM-3007 LoadRDF doesn’t start in parallel mode when sameAs is enabled. This is disabled because the ‘sameAs’ optimisation is not supported in parallel mode and would create inconsistent repositories.
- OWLIM-3039 LoadRDF fails if using an empty ruleset in parallel mode when the input is a zip file.
GraphDB Workbench¶
Bug fixes¶
- WB-998 In class hierarchy diagram - classes with no parents are only shown when there aren’t any classes with parents.
- WB-995 In the SPARQL View, when using a link for a saved query, the execute parameter in the URL (&execute=true) was not handled.
- WB-991 In the Context view - clear button (trash icon) is visible for read-only users. This type of users shouldn’t be able to perform ‘clear’ operation, respectively to see the button.
- WB-990 Read-only users can delete saved queries. Now this is not possible.
- WB-974 Import page hangs when importing many URLs in the Remote Content tab.
- WB-957 In the SPARQL View - after executing an ASK query and then a SELECT query, the default result view is ‘raw response’. The Table view should be default for the SELECT queries.
Improvements¶
WB-899 Saved queries in the SPARQL view are now ordered alphabetically. WB-956 In the SPARQL view, when performing ASK queries the result for boolean values is now True/False, not JSON.
GraphDB Connectors¶
No changes.
GraphDB 7.0.1¶
Released: 27 April, 2016
Component versions¶
Sesame | Connectors |
2.8.11 | 5.0.1 |
Important
This is a patch release, which fixes errors in the packaging:
- Missing Lucene plugin for full-text search;
- Incomplete examples in the Windows and Linux native apps.
The release also provides fixes for some non-critical bugs in the Workbench and the Connectors.
The built-in Java in the native applications has been updated to version 8u92.
GraphDB Distribution¶
Bug fixes¶
- GDB-138 Legacy Lucene plugin not packaged in the distribution.
- GDB-133 Windows and Linux native apps packaged with an old version of the examples.
Improvements¶
- GDB-139 Updated built-in Java to version 8u92 in the native apps.
GraphDB Engine¶
No changes.
GraphDB 7.0.0¶
Released: 18 April, 2016
Component versions¶
Sesame | Connectors |
2.8.11 | 5.0.0 |
Important
- Support for RDF 1.1 providing access to a huge list of public datasets. GraphDB is now fully compatible with all of the latest RDF standards.
- Visual exploration of the imported data sets through the ‘Class hierarchy’ and ‘Class relationships’ views in the GraphDB Workbench. They allow thorough exploration and analysis of the imported data.
- Integrated resource autocompletion that makes it easier to write and edit SPARQL queries in the GraphDB Workbench. Any resource URI can be autocompleted in the SPARQL editor.
- The connectors have been upgraded to new major versions, 5.5.0 for Lucene and Solr, and 2.3.1 for Elasticsearch. In addition to that, the entity filter has been extended to help mapping messy real-world data.
- All JMX features are now exposed over HTTP via Jolokia (in parallel to the jConsole client). It is now possible to manage the server and all other functionality by using cURL. Check the section how to generate a security key in the
./graphdb
startup file. - Health checks to monitor the status of the database - with a single HTTP request it’s possible to check the overall database status and identify if the database is operational.
- Three installation types:
- Desktop installer (new) - for users who are not familiar with Java. They can simply download the installer for their OS and go through the installation steps;
- Stand alone server (new) - a pre-configured web server, which can be started with a simple click and can run in a daemon mode;
- WAR files (deprecated) - the old way of installing GraphDB by using a third-party servlet container.
- Improved and unified database logging - All logs are much simpler, easier to configure and stored in a single directory.
- Upgrade to the latest Sesame version 2.8.x.
- New colour scheme for the GraphDB Workbench that resolves issue with orange text/orange links.
- Java 8 is now required to run the database - make sure to update all your servers to Java 8.
GraphDB Engine¶
New features¶
- OWLIM-1322 RDF1.1 compliance;
- OWLIM-2876 Change the distribution to include an embedded Tomcat. Deployment using
.war
file is still possible, but not preferred; - OWLIM-2897 Reusing a var name from an outer scope in a sub-SELECT should result in an error (configurable);
- OWLIM-2235 Implement health check calls;
- OWLIM-2728 Implement standard query log format;
- The log entry contains all query metadata: query string, bindings, include-inferred, etc.;
- OWLIM-2684 Add Jolokia as part of the distribution;
- OWLIM-2736 Implement extended logging for workers and the standard edition;
- OWLIM-2121 Standalone workers and the standard edition should provide the same logging capabilities as the master;
- OWLIM-2880 The GraphDB engine should support license setting at runtime;
Improvements¶
- OWLIM-2919 License management improvements;
- OWLIM-2927 Update GraphDB default configuration;
- OWLIM-2258 Properly report problems with model conversion through a checked exception; Model conversion errors are now properly logged, which gives a clear picture whether the Sesame’s optimiser is used instead of the GraphDB’s, which is faster;
- OWLIM-2237 Custom ruleset should be copied into the repository image and used from there; It is done on creating of new repo or on adding of a custom ruleset. This prevents the failure of the repository to be instantiated because the ruleset path is pointing to a deleted or a missing tmp directory;
- OWLIM-2079 Disallow adding statements to system graphs;
- OWLIM-2184 Allow to override the default logging configuration with a command line argument and JMX;
- OWLIM-2965 Improve the logging in the cluster code for better errors diagnostic
- OWLIM-2498 GraphDB’s repository
config.ttl
could contain both capital and lower case property values; - OWLIM-2691 Remove RDFPriming from the plugins that are enabled by default;
- OWLIM-2900 Rebuild Entity hash table on startup due to the change in Literal’s hash code in Sesame 2.8.x vs. Sesame 2.7.x;
- OWLIM-2957 LoadRDF should work only in serial and parallel mode.
There are only two modes for adding and inference now - serial and parallel (previously fullyparalel).
Also, the LoadRDF script now uses POSIX syntax for options and parameters. The new implementation registers the created repository in the default SYSTEM repository and adds a new option
-f
to overwrite an existing repository definition; - OWLIM-2981 LoadRDF should work with the existing repository definition.
There is a new option
--id
that allows the usage of existing repository definition; - OWLIM-2964 The SailIterationMonitor mbean in JMX should use the repository id instead of the path to the repository storage. Backslashes in Windows paths can lead to many kinds of problems.
Bugfixes¶
- OWLIM-2973
INSERT
without closing } causes an endless loop. The engine eventually runs out of memory and crashes; - OWLIM-2707 CONSTRUCT query with the same URI for subject and object produces blank node;
- OWLIM-2672 Queries of type
DELETE
...INSERT
...WHERE
ignore the includeinfer
parameter Wheninfer
is set to false inferred statements should not be included in the result of GET requests, but they are still present; - OWLIM-2664 Adding statements to the default graph via the Graph Store protocol fails;
- OWLIM-1677 Unable to abort aggregate queries;
- OWLIM-2622 Queries that generate a lot of new entities at query time result in an exception This can be caused by a query with multiple replace functions on a big repo (> 100M statements). New entities will be generated at query time and if their number exceeds the limit of 1 000 000 an exception will be thrown;
- OWLIM-2160 count(distinct *) counts duplicated rows;
- OWLIM-2962 StorageTool after rebuilding the POS index increases its size instead of decreasing (compacting) it. This happens only in repos with 40 bit IDs;
- OWLIM-2238 NQuadsSimpleParser does not handle baseUri properly. This can happen when trying to import
.nt
or.nq
file that contain relative URIs like <title>, with LoadRDF or GraphDB workbench; - OWLIM-2960 On errors the transaction log will sometimes not rollback the tracker’s information resulting in inconsistent state;
- OWLIM-1738 After replication the rulesets status has been deleted. When using custom ruleset in a cluster - after replication the rulesets status gets lost - the custom ruleset is not default any more;
- OWLIM-2971 After detaching a worker the cluster gets into read-only state;
- OWLIM-2980 INSERT followed by
DELETE
...WHERE
fails in transactional entity pool.
GraphDB Workbench¶
Bug fixes¶
- WB-953 Use daemon threads for import (resolves hang on shutdown in Tomcat)
- WB-934 URIs that contain single quote generate broken links;
- WB-919 Predicate Lists do not show correctly in the Edit Repository form;
- WB-910 Broken link in export view;
- WB-897 Unable to change Use predicate indices and Total cache memory in one step.
Improvements¶
- WB-931 Impose an absolute timeout on the automatic count queries in the SPARQL editor;
- WB-922 New colour scheme for GraphDB 7;
- WB-873 “Download as” in SPARQL view now shows warning when downloaded data might be from a different query;
- WB-861 URI autocompletion in the View resource view;
- WB-845 Add info buttons next to the title of every Workbench view;
- WB-810 Class relationship visualisation;
- WB-756 Autocompletion of URIs in the SPARQL editor;
- WB-730 Class hierarchy and Domain-Range graph visualisation.
GraphDB Connectors¶
Improvements¶
- STORE-229 Upgrade connectors to new major versions of Lucene/Solr/Elasticsearch (5.5.0/5.5.0/2.3.1);
- STORE-233 Implement isIRI/isLiteral/exists in the entity filter;
- STORE-234 Make the entity filter keywords case insensitive;
- STORE-235 Add explicit
xsd:string
support to the Lucene connector (helps with mapping messy data).