Cloud Service: deployed March 20, 2017 (EU, US)
- Concept-map improvements and bug fixes.
- List words fixed limit and total count.
- Consistent results are returned in case of parallel updates and searches.
- Fix SELECT on documents containing more than 1024 child.
- Improve info-words performance when providing multiple words.
- Fix merging progress info in status.
- Improve index recovery on shutdown/start.
- Similar command implemented.
Cloud Service: deployed February 15, 2017 (EU, US)
- Info-words and concept-map command.
- Improved locks for mixed workloads.
- Facet/bimap improvements.
- Internal optimizations and fixes.
Cloud Service: deployed February 1, 2017 (EU, US)
- Improved propagation.
- Use read only copy index for search.
- Added licensing (on-premises).
- Allow timestamps to be used as datetimes.
- DATE more robust.
- Fix density calculation for full text search in JS/SQL.
- Stability fixes.
- Internal code refactoring.
Cloud Service: deployed November 23, 2016 (EU, UK, US)
- Order by date performance improvements.
- Consistently escape select expressions.
- Fix xml to json conversion with invalid tag names.
- Fix transactional search.
Cloud Service: deployed November 9, 2016 (EU), November 4, 2016 (UK, US)
On-premises packages: released November 17 , 2016 (Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7, Debian 7, Debian 8)
- Improved full document access.
- DateTime optimization.
- List empty databases.
- Support ssl in cps-http.
- Support CONTAINS/CONTAINS_FUZZY on joined collections.
- Improved describe-collection output.
- Various optimizations and stability fixes.
Cloud Service: deployed September 22, 2016 (EU), September 23, 2016 (UK, US)
On-premises packages: released October 13 , 2016 (Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7, Debian 7, Debian 8)
- This version contains mostly various bug fixes and stability improvements.
Cloud Service: deployed September 8, 2016 (EU, UK, US)
On-premises packages: released September 13 , 2016 (Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7, Debian 7, Debian 8)
- Correctly escape aggregation values in v3 when response format is JSON.
- Return error when trying to order by non-existent value in JS/SQL.
- Various reliability and performance improvements.
Cloud Service: deployed August 24, 2016 (EU, UK, US)
On-premises packages: released August 26 , 2016 (Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7, Debian 7, Debian 8)
- DDL API release (database management with JS/SQL).
- Datetime support.
- Backup & Restore features release.
- List Words & Alternatives features release.
- Binary data API.
- Improved mixed document handling in search.
- Improved JS/SQL parsing.
- A lot of fixes and improvements on stability and performance.
Cloud Service: deployed July 6, 2016 (EU, UK, US)
On-premises packages: released July 11, 2016 (Debian 7, Debian 8, Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7)
- Group by optimization for string values.
- Add js:type for all strings.
- Obtain shared lock JS/SQL search.
- Reject documents with too long ids.
- Improve snippeting algorithm when text contains very long sentences.
- Fix background reindexing.
- Fix loop in numeric iterators.
Cloud Service: deployed June 14 (EU, US, UK)
On-premises packages: released June 13, 2016 (Debian 7, Debian 8, Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7)
- improved reindexing and recovering from crashes.
- performance improvements of ORDER BY.
- performance improvements of boolean search.
- perfromance improvements of snippet/highlight.
- optimize primary key search (_id == "something").
- improved memory usage.
- added --max_workers argument to node & hub to specify worker thread count.
- various other stability fixes and improvements.
Cloud Service: deployed May 3, 2016 (EU), May 3, 2016 (US)
On-premises packages: released May 3, 2016 (Debian 7, Debian 8, Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7)
- Improvements on stability and performance.
- Create collections now tries to wait for collection to actually start before returning.
- Increased fetch-able document size to 256MB.
- Validate requests for valid UTF-8 (invalid requests get rejected).
- Added index-empty for objects.
- Delete a document if indexing fails of specific document.
- Array length fix for arrays containing objects (requires re-index).
- Fix for facet table collision - when field has multiple types.
- Fixed handling of single and double quotes in snippet/highlight functions.
- Perform full document access in search phase when index is not available.
- Fix CONTAINS on field that does not exist.
Cloud Service: deployed April 14, 2016 (EU), April 18, 2016 (US)
On-premises packages: released April 15, 2016 (Debian 7, Debian 8, Ubuntu 14.04, Ubuntu 16.04, Centos 6, Centos 7)
- Fix sysdb config / policy update.
- Fix bucket compression cache concurreny bugs.
- cps-http crash fixes.
Clusterpoint 4.1.2 - 4.1.4
Clusterpoint 4.1.2 and 4.1.3 had minor internal updates.
Clusterpoint 4.1.4 deployed April 12, 2016 (US)
- SELECT * with joins now returns also joined collection document content.
- Improve parsing of init phase in JS/SQL.
- Improve processing of special documents (reindex, snapshot).
- Ignore whitespace between JS/SQL operators.
- Added check for outdated collection config.
- Improve CONTAINS/CONTAINS_FUZZY functions to skip documents, if field's xpath cannot be determined.
- Report index disk usage in request statistics.
- Correctly parse escaped characters when creating WHERE expression tree.
- Created manager commands to edit database slots.
- Discard collection permissions from all users when collections is dropped.
- Backup/Restore with snapshots implementation.
- Added REST v3 standalone configuration.
- Fix bug with snippet/highlight when value contains invalid xml (e.g. html tags).
- Fix nested object handling in fetch phase.
- Fix bug in JS/SQL destination parser.
- Fix local IP for package installs.
- Fix infinite loop in tribool table scan operation when matched document count is over 50K.
- Fix disk space calculation bug that caused hard links to count as new files.
- Fix very small wildcard searches also matching exact match identifiers.
- Fixed manager slave authorization and listing of v4 collections.
Deployed March 31, 2016 (EU) and April 1, 2016 (US)
- Permission check of joined collections.
- Check agains correct ID paths in JOINs.
- Improved lookup table initializing values.
- Throw JS exception when trying to run CONTAINS on fields that do not have text index.
- Do not allow requests to disabled databases.
- Fixed binding and indexing of boolean data type.
- Fixed array ordering.
- Fixed document count errors.
Deployed March 22, 2016 (US) and March 21, 2016 (EU)
- Database -> Collection support. Database consists of Collections, Documents are stored in Collections.
- Array support and array indexing (alows to search in arrays).
- EMBED operator - combines all group values into an array.
- ID field length limitation (up to 249 characters).
- ID field full text indexing.
- Improved string access and storage (now it is faster).
Deployed February 18, 2016 (US & EU)
- In case of 3 or fewer replicas use all replicas for search/update.
- Removed some of locks from get status command, so it returns faster responses.
- Multiple fixes for invalid memory writes/reads.
- Fixed Integer aggregatinon (v3).
- Query optimization fix on full text search in document root.
- Send reindex command in parallel to all shards.
- LEFT JOIN support.
Deployed February 3, 2016 (US) and February 8, 2016 (EU)
- Fixed number/string parsing and conversion when indexing.
- Various index corruption fixes.
Deployed January 20, 2016 (US & EU)
- Support JSON string aliases for select columns in JS/SQL.
- Optimized "SELECT * FROM x LIMIT n" for faster initial load of run query view.
- Support for differencing between missing field and field with empty string value in v4 databases. Enables search for empty values.
- JS/SQL XML support.
- Support arrays, iterators, iterables and generators as results of ORDER BY expressions.
- Reworked datetime functions JS/SQL. Now all date functions return
time since epoch in seconds (Unix timestamp).
Deployed December 16, 2015 (EU), December 17, 2015 (US)
- Extended query optimizer to replace JS/SQL standard aggregation functions with native code.
- Fixed issue that database Clear removed data model overrides.
- Fixed issues with data model persistence.
Deployed December 2, 2015 (EU), December 1, 2015 (US)
- Added JSON parsing error details.
- Fixes in fuzzy search. Some rewrites were not properly discovered.
- Fixed issues with field bindings in SELECT clause.
Deployed November 4, 2015 (US & EU).
- Added fuzzy search in JS/SQL with CONTAINS_FUZZY operator.
- Added INSIDE GeoSpatial functions.
- Added multi-ordering in JS/SQL.
- Added stemming languages in data model.
- Handling case when stored document cannot be parsed to JSON.
Deployed October 29, 2015 (US) and October 30, 2015 (EU).
- Added XML multi-document insert and update.
- JS/SQL biniding of XML documents without root tag.
- Internal cleanup: data model overrides during recreation of database.
- Bugfix: crash in QueryParser::parse_query when given an empty string.
- Bugfix: JSON parsin and array binding issues at fetch stage.
- Bugfix: data model thread safety and recovery.
Deployed October 21, 2015 (US) and October 22, 2015 (EU).
- Added Distance2D and DistanceGeo functions in JS/SQL.
- Added GROUP_KEY function in ORDER clause.
- Compacting of data deleted more than 24h ago.
- Allow ':' in field names.
- Bugfix: data model cardinality.
- Bugfix: return correct hit count when grouping.
Deployed October 16, 2015 (US) and October 15, 2015 (EU).
- Added grouping support by multiple values e.g. "SELECT COUNT() FROM db GROUP BY Array(1,2,3)", which would add every document in 3 groups.
- Added multi-document insert / retrieve.
- Improved memory usage and reporting.
- Improved manager start-up time.
- Added repository compression.
- Fixed multiple bugs in JS/SQL syntax parser.
- Fixed multiple bugs in REST API.
Deployed October 8, 2015 (US & EU).
Added basic support of JS/SQL query language.
The main differences from Clusterpoint 3 query language:
- Native support of JSON and XML data types. Clusterpoint 3 allowed to store JSON documents, but internally converted them to XML. Clusterpoint 4 allows to store every data type natively and retrieve in exact form it was stored.
For every database created in Clusterpoint Cloud user have to specify the version of the database (v3.0 or v4.0).
Clusterpoint 4 currently provides only REST API. Native APIs for other programming languages will be available with the Clusterpoint 4.1 release.