The ‘Long Index edition’ of Sisense determines how Sisense transforms your data when loading it in to an ElastiCube.

Sisense builds your ElastiCube using pointers that represent your data. These pointers are used to find the actual values so no information is lost.

To reduce memory consumption and improve performance, Sisense builds ElastiCubes with 32-bit pointers reducing the size of the pointers.

You may experience “Out of memory” errors when working with big data such as:

  • 300 million rows of numeric fields
  • 2GB of string fields

For example, let’s assume that your data contains strings with a length of 1024 characters and 4,194,304 unique rows such that 1024*4,194,304 > 4GB in content after indexing, you should install the Long Index version of Sisense.

By installing the Long Index edition, Sisense will use 64-bit pointers. While this is fine for big data, you should use the default installation of Sisense when working data smaller than the figures mentioned above for faster performance.