hi,I would like to know what the underlying data structure of the Nonclustered and clustered Columnstore Index (In-Memory) is. I mean the column entries are somehow compressed and loaded into memory in both cases, but how is the data itself structure in the memory? Are there indexes created first from columnvalues, or is there a hashtable being constructed from each of the column values ? and from the columnvalues, how are the rows built together again? I mean are they numbered sequentielly in parallel for each column? so if I ask 20th entry from column x, then it looks automatically for the 20th column entry in y ?
↧