InMemory.Net Feature List
- Pure Dot Net In Memory Database
- Runs on Windows, Mac, Linux
- Easily Import data through ODBC/OLE DB/ Dot Net Data Drivers, through simple text scripts or via a Dot Net API. E.g.
DATASOURCE a1=ODBC ‘dsn=your_DSN’ IMPORT sales=a1.sales SAVE
- 32 Bit / 64 Bit Odbc Drivers
- DISTINCT/MIN/MAX/SUM/AVG/COUNT/COUNT DISTINCT Support
- INNER / LEFT Join Support on multiple fields
- UNION ALL Support
- SQL Mostly Compatible with SQL Server
- Data is stored in Columnar Data Structure. Bit packed on Disk. Lookups expand to Byte/Short/Int in memory.
- Subquery support, with column elimination and parent where clause pass through. NOCACHE keyword to control, whether common suqueries to be turned into tables.
- Currently support 14 of TPC-H queries. Able to execute about 100k queries on 1G TPC-H Dataset in 1 hour ( 4 Core CPU).
- Highly parallel query execution during main stage, reduction and results processing.
- Joins between tables can be created dynamically at Run Time, and are not hard coded during extract phase
- Clustered Index feature.
- Easy to Embed core technology in Dot Net products. ( Currently 2 DLLs )
- COM & Dot Net Api for connecting and communicating with database.
- Queries compiled to Dot Net assemblies, and then JIT compiled to run very quickly. Constants, tables & column names abstracted to allow reuse.