Code: Select all
data_type | encoding_type | count
----------------+------------------+-------
float | AUTO | 2782
float | RLE | 1214
float | BLOCKDICT_COMP | 1100
float | DELTARANGE_COMP | 1026
int | DELTARANGE_COMP | 804
int | RLE | 696
int | AUTO | 454
date | RLE | 360
int | BLOCKDICT_COMP | 328
int | COMMONDELTA_COMP | 292
float | COMMONDELTA_COMP | 286
date | BLOCKDICT_COMP | 158
timestamp | GCDDELTA | 148
timestamp | DELTARANGE_COMP | 140
timestamp | AUTO | 130
timestamp | RLE | 114
float | BLOCK_DICT | 90
date | COMMONDELTA_COMP | 80
timestamp | COMMONDELTA_COMP | 56
date | AUTO | 36
timestamp | BLOCKDICT_COMP | 34
boolean | RLE | 28
int | DELTAVAL | 22
boolean | AUTO | 16
date | DELTAVAL | 14
int | BLOCK_DICT | 10
int | GCDDELTA | 8
numeric(8,3) | DELTARANGE_COMP | 8
numeric(9,3) | COMMONDELTA_COMP | 8
numeric(10,3) | BLOCKDICT_COMP | 8
date | DELTARANGE_COMP | 6
numeric(19,5) | BLOCKDICT_COMP | 4
numeric(37,15) | AUTO | 4
date | BLOCK_DICT | 2
I would like to improve the AUTO encoding to use a more appropriate encoding. Would the query-specific DBD run be the best way to do so?
Cross posted on Vertica Community.