SciDB Release 18.1.15

  • Documentation is online here. Note that some features described are only available in the Enterprise Edition.
  • Installation script and source tarball can be downloaded here.
  • A sample virtual machine can be downloaded here
  • JDBC is available by installing (yum or apt-get) the scidb-18.1-jdbc package.
  • Additional prototypes from Paradigm4 Labs are available on GitHub; these include

Product questions should be directed to the appropriate topic on this forum.

Amazon Machine Images with more example datasets and queries are available upon request. To request AMIs for specific use cases, please email or

Generally, use

Index of SciDB Releases
SciDB community edition installer script

There have been a lot of minor releases for the major release 18.1. The latest is 18.1.10. You can check your version by doing

scidb --version
SciDB Version: 18.1.10
Build Type: RelWithDebInfo
Commit: 6a1e38e
Copyright (C) 2008-2017 SciDB, Inc.

If you have setup your scidb using repositories, you can use

# CE
yum update scidb-18.1 

# EE
yum update paradigm4-18.1 
yum update paradigm4-18.1-p4

… and equivalent commands for apt-get.

If you are using Docker, you would need:

docker pull rvernica/scidb
docker pull rvernica/scidb:18.1
docker pull rvernica/scidb:18.1-ext
# and so on

Among other things, the minor releases from 18.1.1 to 18.1.10 include

  • performance improvements to the query planning and execution engine
  • performance improvements for summarize operator
  • a bug-fix for arrays with compression turned on
  • security namespaces are now first-class names in AFL, so you no longer need to provide them as string literals to the various namespace operators

For the above reasons, upgrading to the latest 18.1 version is highly recommended.

As noted at this post, if you are upgrading from <18.1.7 to greater, you would need to recompile plugins too. A little bit of a pain, but it will be worth it :slight_smile:

We will keep you posted about other minor releases on this thread.

How to using JDBC by SciDB 18.1?

SciDB 18.1.11 is out!

Improvements in this release

This minor release contains significant improvements related to array versioning.

  • Speedup in data access when there are many versions of an array
  • Ability to remove all but N versions of an array by calling remove_versions(ARRAYNAME, keep: N)

remove_versions(A)              # - keep all but the latest version of A
remove_versions(A, v)           # - keep versions older than version v
remove_versions(A, keep: N)     # - keep the last N versions of A (or all of them if there are fewer than N)


SciDB 18.1.13 is out!

Improvements in this release

This release contains improvements under the following headings:

  • SciDB as a service
    • A fix to SciDB service module for RedHat / CentOS 7
    • Renamed SciDB-xx.y service to scidb-xx.y (see note below)
  • SciDB logging
    • Default SciDB log level fixed to INFO (instead of DEBUG).
    • Prevent SciDB upgrade from overwriting log level with default.
  • SciDB security
    • A fix to the security library by which characters like @, . can be used as part of the username. Now, usernames like can be used as SciDB usernames (this was not the case earlier)

Note about renaming of SciDB service

There is no corrective code to remove or rename the old service (SciDB-xx.y) so you must, before installing the upgrade, follow the instructions in the manual to remove the service. Otherwise there will be a SciDB-18.1 service AND a scidb-18.1 service (THIS APPLIES TO ALL PLATFORMS: centos 6 and 7, rhel 6 and 7, ubuntu 14.04).


SciDB 18.1.15 is out!

Improvements in this release

  • Fixed: remove becomes very slow when there are many chunks
  • Fixed: bug when the same array name is used in different name spaces
  • Fixed: storage manager does not return space freed by delete_versions to the filesystem.
  • Fixed: show_role_permissions with null argument can potentially bring SciDB instance down
  • Implemented operator _broom() within which various SciDB cleanup tasks may be performed, such as forcing a RocksDB full compaction.