Neo4j pre 1.9 Advanced changelog

1.8.RC1 (2012-09-05)

Kernel: o Removed contention around allocating and moving persistence windows so that a thread won't need to await another thread doing this refresh, instead just continue knowing that the windows will be optimally placed in a very near future. o Removed contention around translating a key (as String) into the ID by using copy-on-write map instead of a concurrent hash map. Used in property key as well as relationship type translation. o Fix for Node/Relationship#getPropertyValues() sometimes returning null values from the iterator.

Server: o Upgraded Jackson JAXRS to version 1.9.7 o Keeping the Cypher execution engine between calls makes it possible to re-use execution plans o added User-Agent header tracking to udc to determine rest-driver usage

Cypher: o Removed the /../ literal for regular expressions. Now a normal string literal is used instead o Concatenation handles types better o Changed how the graph-matching module is used, to make it safe for concurrent use o Better type error messages o Renamed iterable to collection o Fixed #795: so that WITH keeps parameters also for empty aggregation results o Fixed #772: Creating nodes/relationships goes before SET, so SET can run on already created elements o Added error when using != instead of <> o Fixed #787: Issue when comparing array properties with literal collections o Fixed #751: Better error message for some type errors o Fixed #818: Problem where properties could only have scalar values (they can be arrays as well) o Fixed #834: Gives relevant syntax exception when properties are entered in MATCH

1.8.M07 (2012-08-08)

Kernel: o Traversal framework backwards compatibility

• Cleaned up any breaking changes
• Removed Expander#addFilter o Kernel JMX bean instance identifier is now reused and can optionally be set explicitly via forced_kernel_id config setting

Server: o Consoles in webadmin can now be disabled.

Cypher: o Added escape characters to string literals o Renamed RELATE to CREATE UNIQUE

UDC: o Added edition information (community, advanced, enterprise) o Added a cluster-name hash so that stores originating from the same cluster can be aggregated o Fixed release version and revision o Changed precedence of database configuration over internal udc configuration o Added distribution information (dpkg, rpm, unknown)

1.8.M06 (2012-07-06)

Kernel: o Deprecated AbstractGraphDatabase.transactionRunning() o Changed synchronization of applying transactions to prevent a deadlock scenario o Original cause can be extracted from a transaction RollbackException

Server: o Fixed issue that stopped the server from starting without the UDC-jars.

Cypher: o Fixes problem when graph elements are deleted multiple times in the same query o Fixes #625: Values passed through WITH have the wrong type o Fixes #654: Some paths are returned the wrong way

1.8.M05 (2012-06-25)

Kernel: o Configurable amount of logical logs to keep, by for example number of days or size on disk. keep_logical_logs configuration is extended to support values such as: "10 days", "200M size" a.s.o. Regardless of configuration there will always be at least the latest non-empty logical log left. o Reduced synchronization while memory mapping files, leading to increased multithreaded performance

Server: o Added support for multi line Cypher queries in webadmin browser

Lucene-index:

• Removed lucene_writers_cache_size and have lucene_searcher_cache_size decide for both, since it's doesn't make sense to have a writer w/o a searcher and isn't possible to have a searcher w/o its writer.
• Loosened contention regarding getting index searchers for querying.

Cypher: o CREATE and RELATE can now introduce path identifiers o String literals can now contain escape characters o Fixes #600: Double optional with no matching relationships returns too many rows o Fixes #613: Missing dependencies not reported correctly for queries with RELATE/SET/DELETE o Fixes around optional paths

1.8.M04 (2012-06-07)

Kernel: o Additions to the traversal framework:

• Bidirectional traversals (with BidirectionalTraversalDescription). AllPaths/AllSimplePaths uses this feature and are now faster due to less relationships being traversed.
• Multiple start nodes (as well as multiple end nodes for bidirectional traversals).
• PathExpander (RelationshipExpander for Paths) which has the full Path passed in instead of just the end node of the path. It can also see and modify a user defined traversal branch state.
• Metadata about the traversal, for the time being: number of relationships traversal and number paths returned.
• Added Path#reverseNodes() and Path#reverseRelationships which gives all nodes/relationships in a path in reverse order (starting from the end node going back to the start node). More relevant in many scenarios as well as allowing for a more efficient implementation.
• Sorting of traversal results, specify with TraversalDescription#sort(Comparable).
• Some measure to reduce memory consumption and garbage generation during traversals.

Graph-algo: o AllPaths/AllSimplePaths uses the new bidirectional traversal feature in the traversal framework. Less relationships now needs to be traversed to find the requested paths. o Added an implementation of the shortest path algorithm with the bidirectional traversal feature.

Cypher: o CREATE now accepts full paths and not only single nodes and relationships o Path patterns can now be used as expressions and as predicates o Fixes bug where RELATE fails to compare array properties o Fixes #573: Arithmetics operations have wrong type o Fixes #567: Parameter maps coming through REST are not parsed correctly o Fixes #563: COUNT(*) does not return 0 for empty result o Fixes #561: RELATE doesn't work as expected with map-parameters

1.8.M03 (2012-05-24)

Kernel: o Changed array map to CHM in property index manager, better multithreaded performance Shell: o Added BEGIN TRANSACTION/COMMIT/ROLLBACK o Sessions now live on the server side instead of on the client, which means that not just serializable values can be set there. Paves the way for Cypher making use of it. Server: o keep_logical_logs is now respected if set to false - default is still true Cypher: o Added RELATE o Changed the CREATE syntax, so it looks more like RELATE o Fixes #506: delete head(rels) doesn't work o Fixes #508: length function on array properties doesn't work o Fixes #512: start n=node(*) match n-[r?]->() delete n,r not working o Fixes #514: Extract doesn't work with array properties o Fixes #523: String literals not parsed correctly in CREATE statement o Fixes #526: cypher filter in return clause o Fixes #536: SQRT only returns an integer o Fixes #543: Appending elements to collections

1.8.M02 (2012-05-11)

Kernel: o Optimized short string and short array encoding algorithms o Fixed problem with SpringTransactionManager during HA master switch Shell: o Shell can now be exited with Ctrl-D Server: o Support for streaming results for http batch operations. o Proper encoding of Index URI key/value pairs. o Fixed Swedish character problem in webadmin string properties. o Webadmin chart labels have nice formatting now. Cypher: o Added the possibility to create nodes from an iterable of maps

1.8.M01 (2012-04-26)

• Byte array properties are handled in a more optimized way, increasing performance by a couple of times at least.
• Fix for an issue where update of cache when creating a relationship on a very dense node would take longer and longer time.
• Fix for an issue where a recovery would sometimes fail with it reporting that a start record already had been injected.
• Added "create" shell app for Cypher queries only doing creations. Server: o Added streaming to REST API. All representation implementations have been adapted to produce their data lazily. o Added HTTP logging. Cypher: o Added the possibility to return all graph elements using RETURN * o Added CREATE, SET, DELETE, FOREACH, WITH o Fixes issue #432, #446

1.7 (2012-04-18)

• Moved BatchInserter to a different package.
• Fixed 'Record Not In Use' issue with creating and setting Node properties with the BatchInserter. Cypher: o Added the possibility to use multiple relationship types on patterns o Fixed bug when aggregation is used in ORDER BY and not listed in the RETURN clause Server: o Added scored index results to REST API o Improvements to installation docs for the server o Added auto index management API to REST, fixing #399 o Fixed unicode issues in batch operations API in windows and OS X o Server now disallows creating indexes with empty names. Closes #311 o Attempting to delete a non-existing index now returns 404, closes #349 UDC: o Modified UDC to run one thread per DB, and to shut down background thread on unload. Closes #279 o UDC now sends the machines MAC adress to Neo Technology, to separate between instances behind firewalls o UDC now sends database registration ID to Neo Technology, if one is available o UDC now sends "tags" to Neo Technology, these contain information about the type of deployment, such as language, web-container, app-container, spring, ejb

1.7.M03 (2012-04-11)

• Removed old OSGi support in favor of the new and better one.
• Added possibility to use logback for logging.
• Renamed array cache to GC resistant cache (GCR) and moved it to enterprise.
• Fixed problem with GCR cache that could cause it to calculate incorrect size.
• Fixed problem with closing of messages.log on windows

1.7.M02 (2012-03-26)

• Added lock free atomic array cache that avoids triggering full GCs.
• Added GC monitor that will keep track on how much the JVM blocks application threads.
• Fix for an issue where upgrading indices with an unclean shutdown.

1.6.M03 (2012-01-12)

• Added the ability to monitor the locks (optionally filter on contended-only and how long they've lived) in the LockManager through JMX.
• Added a management bean for the diagnostics logging SPI.
• Adding possiblity to restart server via JMX.

1.6.M02 (2011-12-16)

• [cypher] BREAKING CHANGE: Changed the syntax for iterable predicated ALL/NONE/ANY/SINGLE to use WHERE instead of a colon
• [cypher] BREAKING CHANGE: ExecutionResult is now a read-once, forward only iterator.
• [server] DEPRECATION: Cypher execution is now part of the core REST API, the cypher plugin is deprecated.
• [server] Webadmin data browser now supports cypher queries
• [server] Updated to gremlin 1.4
• [server] Fixed bug in how auto indexes are represented
• [server] Max request execution time can now be limited
• [server] Hypermedia URLs returned by the server now set their host based on the HTTP host header
• [kernel] Lower memory usage of ImpermanentGraphDatabase.
• [kernel] Abstracted stores that stores id/name pairs into AbstractNameStore/AbstractNameRecord and removed lots of duplicated code.
• [kernel] Fixed a race condition in the property cache which cuold poison the cache.
• [kernel] Fixed an issue where a JVM crash in the wrong place would make the next startup rename the current logical log without incrementing the logVersion, making the next log rename fail.
• [kernel] Start records in the logical log aren't written until the transaction starts preparing/committing. This fixes an issue with the logical log not being rotated sometimes.
• [kernel] Added AbstractGraphDatabase#tx() which returns a TransactionBuilder possible of configuring the transaction to be "unforced" meaning that logical logs won't be forced when committing. Means better performance for small non-critical transactions.
• [kernel] Reduced number of proxy object instantiation and node record loading in general.
• [kernel] Added a wrapping graphdb abstraction, which makes instances survive even if the underlying db restarts.
• [cypher] Added allShortestPaths
• [cypher] Added COALESCE
• [cypher] Added column aliasing with AS
• [cypher] Fixed problems with optional graph elements

1.6.M01 (2011-11-24)

• Added support for a branched data bean (implemented in HA)
• Added info on latest committed txId in TransactionManager bean

1.5 (2011-10-31)

• Added default remote JMX access configuration and security setup options

1.4.M04 (2011-06-09)

• First iteration of the Cypher Query Language included, with a section in the manual.
• Experimental support for batch operations in REST
• The Neo4j Manual now includes some examples, and the section on REST has been extended.

1.4.M01 (2011-04-28)

• Server logging has been changed, see the Server Configuration chapter in the manual for further information.

1.3 (2011-04-11)

• Neo4j Community now licensed under GPL
• All known Windows problems in Neo4j fixed.
• rewritten Webadmin interface with graph visualization support
• Short string support in kernel for better
• 256 Billion database primitives
• support for Gremlin 0.9 and related Tinkerpop projects
• Better JMX and monitoring support (Neo4j Advanced)
• New backup solution with support

1.3.M05 (2011-03-24)

• New look and feel of the webadmin
• New visual data browser
• Updated to Gremlin 0.8
• Added docs for server plugin initialization.

1.3.M04 (2011-03-10)

• All manpages are included in the manual.
• Some fixes to be more Windows friendly.
• Added Dijkstra to the list of graph algorithms to be used when finding paths.
• Support for advanced index queries through REST.

1.3.M03 (2011-02-24)

• Numerous updates to the Webadmin tool to make it more usable and to fix some visual bugs.
• Removal of the properties service, replaced with a discovery service to allow third-party apps to bind to the server at runtime and discover what services are offered and where.
• Changed the configuration file format to make it simpler and less error prone (but it is not) backwards compatible with earlier versions.
• Separated out the server plugin API for easier development, making only the only development dependency for server a very thin layer and small jar.

1.3.M02 (2011-02-10)

• Gremlin updated to 0.7 and trimmed down to bare bones dependencies.
• Webadmin minor improvements in Console, Data browser and monitoring.
• better REST JSON property support in Arrays.
• server started in High Availability mode.

1.3.M01 (2011-01-27)

• Full online backup (no need to copy store files before hand) in HA and general support for requests with unlimited size.
• A couple of bugs fixed in lucene index.
• More memory efficient handling of transaction streams (extracting and applying).
• New 'eval' shell command which lets you execute JavaScript on the database.

• Server

• Ability to start the server in HA mode.
• Excludes the neo4j-index component (not used anyhow) to enable HA mode.

• Examples

• Added an example of an ordered traversal.

1.2 (2010-12-29)

• New components:

• Neo4j Server, including Web Admin
• Neo4j High Availability
• Neo4j Graph Database Monitoring and Management tools moved to its own component
• Neo4j Index API integrated in the GraphDatabaseService API
• Neo4j Usage Data Collection

• Other changes:

• Additional services (extensions) for the Neo4j kernel are loaded automatically or can be injected in a running instance.

• Improved memory footprint and read performance.

• A weak reference cache is now available for high load low latency workloads.

• The old index API has been deprecated (but still included and have been updated to use Lucene version 3.0.1).

• There is a new index API based on Lucene supporting multiple indexes both for nodes and relationships.

• Path algos can now be performed using the shell.

• All known bugs have been fixed. For more details see the individual milestone releases below.

1.2.M06 (2010-12-21)

• Kernel

• Fixed an issues with PruneEvaluators getting invoked with the start node as argument. PruneEvaluators was never invoked with the start node before the arrival of Evaluator interface.
• Added logging of recovery completion.
• Added TransactionData#isDeleted(Node) and #isDeleted(Relationship) and also made the collection to hold these a HashSet instead of ArrayList. This allows fast lookups for asking if a node/relationship is deleted in a transaction.
• More flexible caching (can instantiate an arbitrary Cache object).

• Shell

• Fixed a problem where ShellServerExtension assumed the shell was enabled during shutdown, which could result in NPE.

• Lucene-index

• More flexible caching (can instantiate an arbitrary Cache object).
• Merged the fast-writes branch. Basically it makes committing changes to lucene indexes faster and more throughput-friendly for multiple threads. The performance improvement for committing lucene transactions shows best for small transactions, but will improve all modifying operations to lucene indexes.
• Fixed bug which could close an IndexSearcher before it was supposed to when there was exactly one item in the result set, two or more threads did this at the same time as there was another thread doing an update... AND the reading threads all called IndexHits#getSingle() on the result.

• HA

• When a new master is chosen, slaves throws away the id allocations they got from the previous master. This fixes a problem where occupied ids where reused.
• Enabled (and fixed problem with) thread on master which checks for dead slave connections and rolls back those transactions, so that their locks are released. This fixes a problem where an HA cluster would seem to freeze up and become unresponsive to write requests.
• Adding Log4j and Netty license information.
• Also closes the executor containing the connections.
• Changed dependency scopes.
• Shuts down databases after verifying them.

• Server

• Bridge OSGi LogService to server Logger. Bundle log messages and stdout pipe through correctly.
• Refactoring of functional tests to remove static dependencies. Introduced a server builder to deal with it instead.
• Can now add performance-tweaking properties to the database hosted by the server. It uses the same neo4j.properties file as the embedded version, but you need to add a property: org.neo4j.server.db.tuning.properties into the neo4j-server.properties file.
• Webadmin: Removed component titles, added save button to data browser, minor change to data browser layout.
• Webadmin: Minor UI updates, added error message that shows up when server connection is lost.
• Webadmin: Moved charts into tabbed box, minor UI updates.
• Fixed an issue with initialization order of things in the server to make sure that RRD is initialized when it is needed.
• Webadmin: Added proper tab-like styling to chart tabs and scale selectors.
• Fixed bug in RRD memory usage sampler, making rounding happen at end of calculations, instead of in the midst of.
• Webadmin: Made 30minutes the default view in charts.
• Webadmin: Added kernel version to dashboard, fixed bug in chart drawing (was drawing when the charts are not visible under certain conditions).
• Webadmin: Disallow self relationships in UI, and show error message explaining this.
• Can now load 3rd party JAX-RS jars from the classpath.
• Updated static assembly to properly include both webadmin statics and documentation. Documentation now tested and ships properly down to neo4j-standalone.
• Webadmin: Made the current node show up by default when adding new relations in webadmin.
• Webadmin: Dashboard shows charts with tabs, and allows switching between charts.
• Added a basic readme with instructions for building, running, and functional testing.
• Updating the way to create temporary files to solve the functional test problem where lots of files stick around.
• re-implemented REST to expose indexing of nodes and relationships (new index API)
• single path algo works in RestfulGraphDatabase now
• fixed duplicate paths for delete node or relationship from index
• Http DELETE requires the proper mediatype to accept
• All indexes must be created to use them.
• Added the ability extend the REST API with server plugins.
• Added back links to the first page in the HTML format.
• updating to Gremlin 0.6
• Updated the component site documentation with the new index API.

1.2.M05 (2010-12-02)

-New components

• neo4j-ha, providing high availability

-Important changes, server

• updated to Jersey 1.4
• integrated index is now supported by the shell
• new Evaluator interface improves the traversal API
• support for index hit scores in neo4j-lucene-index
• index caching support added to BatchInserter

1.2.M04 (2010-11-18)

• added a server packaging as part of neo4j
• added more configuration options to neo4j-lucene-index
• generating javadocs for all components automatically to components.neo4j.org/{component-name}
• Added ImpermanentGraphDatabase to aid in testing

1.2.M03 (2010-11-04)

• Monitoring/management over JMX was moved to the new neo4j-management component.
• Added ability to get the name and configuration of existing integrated indexes.

1.2.M02 (2010-10-21)

-New components

• Lucene index, an implementation of the integrated index framework using lucene as backend. Planned to replace current index service.

-Important changes

• Fixed shutdown issue regarding UDC leaving a thread running.
• A new index framework integrated into the kernel API.
• Getting relationships is faster due to less cache lookups.

1.2.M01 (2010-10-08)

-New components

• Udc

-Important changes

• Index:
  • Lucene version upgraded to 3.0.1
• Shell:
  • Apps discoverable with service locator
  • Added read-only mode on server
  • 'host' and 'config' options
• Graph-algo:
  • Find paths of certain length
• Kernel:
  • Lower memory footprint
  • Higher throughput/concurrency for reads
  • Common interface for loading kernel extensions (f.ex. shell)

1.1 (2010-07-29)

-New components

• Graph-algo
• Online-backup

1.0 (2010-02-16)

-Initial components

• Kernel
• Index
• Remote-graphdb
• Shell