What's New?
Version 4.7
- Added support for Window functions in OmniSciDB. This is an early iteration of an often-requested SQL feature allowing for analytic computations over rolling windows. See the release notes and documentation for details and known limitations.
- Added support for Visual Data Fusion, the ability to add measures on a combo chart from more than one source table, and also crossfilter across these sources. This is a natural complement to multi-layer geospatial chart.
- Distributed Rendering support for polygons supporting much larger shape datasets within analytical workflows.
- Filters on aggregates in Pointmap and Scatter Plot, equivalent to applying a HAVING filter on SQL aggregate queries.
- INSERT TABLE AS SELECT (ITAS) support in both single-node and Distributed modes, allowing for easier ELT (Extract/Load/Transform) type workflows in OmniSciDB, typically combined with CTAS (CREATE TABLE AS SELECT).
- Data import status within Immerse for time estimation and import success information.
Read more at the OmniSci Blog
Version 4.6
- CTAS (CREATE TABLE AS SELECT) on distributed installations.
- Import Parquet format data files.
- Updates on variable length columns.
- Single Sign-on with SAML for compatibility with Okta and other providers.
- Extended NULL support for variable length arrays. The full array can now be null, in addition to individual elements.
- Support for high-precision timestamps, up to nanosecond precision.
- Improved performance loading String Dictionary from storage.
- Significantly more scalable rendering from projection (non-aggregate) queries.
- Immerse Data Manager:
- Delete, Append, and Truncate (delete all rows from) tables.
- Auto-detect header rows on import, manually indicate whether the first row is a header row.
- Immerse:
- Improved performance for non-aggregated Choropleth/Pointmap/Scatterplot charts.
- Non-aggregated Choropleth now cross-filters on zoom.
Read more at the OmniSci Blog
Version 4.5
- Enterprise trial version is now available.
- Better memory handling through improved estimation of GPU memory requirements. Automatically run query on CPU if not enough GPU memory is estimated to be available.
- Better handling of NULL values.
- DECIMAL/NUMERIC fields can be downcast to different scales and precisions.
- Dictionary size increased to 2.15 billion entries.
- Add support for the lasso filter on Linemap chart.
- Added clarity to formatting options and created a new option to represent billions as B.
- Default ports changed from 9090-9094 to 6273-6280 to avoid collisions.
- Renamed key components from MapD to OmniSci. See the Release Notes.
Version 4.4
- Improved geospatial function support
- Support for pct/blend accumulation rendering modes in distributed configurations
- Improved error tracking
- Support for SAML authentication with Okta
- Improved performance on String Dictionary import for multiple String Dictionary-encoded columns
Version 4.3
- More robust joins between different types
- Better compression on decimal/numeric types
- More efficient rendering of lines using the GPU rather than first copying results to the CPU
Version 4.2
- Linemap Chart
- Percentage (relative) views in Stacked Bar and Combo charts
- Top-level ACCESS privilege
- Faster imports of dictionary-encoded string data
Version 4.1
- Stacked Bar Chart
- ALTER TABLE ADD COLUMN is now supported.
- GRANT/REVOKE ALL ON DATABASE are now supported.
Version 4.0
Read more at the OmniSci Blog