These release notes apply to version 19.11 of SciDB; they cover all features and changes since version 19.3 of:

• SciDB Enterprise Edition (EE)
• SciDB Community Edition (CE)

Supported Operating Systems

SciDB 19.11 supports the following platforms:

• Ubuntu 16.04 (new for Release 19.11)
• Ubuntu 14.04
• RHEL 7

• CentOS 7

Support for CentOS 6 has been discontinued in this release.

SciDB Features and Changes

Data Structures

• Data Frames now support replicated distributions.
• Arrays may now be created with an option to compress the empty bitmap which can result in significant storage and memory savings.

Operators

• lock_arrays() - creates a global array lock that prevents new arrays from being created and prevents updates to existing arrays.
• mquery() - support for mult-query which ties the transactionality of multiple queries together.  If any query in the multi-query failed, the system rolls back to the state before any of the queries was executed.

Performance and Resource Usage Enhancements

• Improved performance in join(), cross_join(), store(), load(), and save() operations.
• Improved memory usage in sort() operations.

Known Issues, Limitations, Behavior Changes

• The storage format has changed.  If upgrading to the current release using data previously created, a new table must be created before SciDB is started.  The table should be created as follows:

Enter the postgres shell:
$ sudo su postgres
$ psql mydb

Issue the following two postgres queries:

create table "provisional_arrays"
(
  array_id bigint,
  array_version_id bigint,
  array_version bigint,
  instance_id bigint references "instance" (instance_id),
  multiquery_id varchar,
  unique(array_id, array_version_id, array_version, instance_id)
);

alter table provisional_arrays owner to scidb_pg_user;

Paradigm4 Labs Features and Changes

•  Paradigm4 provides a host of external tools and add-ons in the form of GitHub repositories at http://github.com/paradigm4. Please contact us for guidance on using these.