redshift best suggested recommended column encoding by data types and by encoding types. column encodings: raw AZ64 Byte-dictionary Delta LZO Mostly Runlength Text255 ZSTD Data types:BOOLEAN DOUBLE SMALLINT INTEGER BIGINT DECIMAL REAL DOUBLE CHAR VARCHAR DATE TIMESTAMP TIMESTAMPTZ TIME

distribution style.md

style	use case
key	same value same slice, for joins
all	dimension small table for frequent joins, data is copied to all nodes first slice
even	unknown cases

• Set DISTKEY to the column most used in a JOIN
• Set SORTKEY to the column(s) most used in a WHERE

redshift column encoding.md

column type	encoding
BOOLEAN	raw
DOUBLE PRECISION	raw
SMALLINT	AZ64, ZSTD, delta, mostly
INTEGER	AZ64, ZSTD, delta, mostly
BIGINT	AZ64, ZSTD, delta, mostly
DECIMAL	AZ64, ZSTD, delta, mostly
REAL	raw
DOUBLE PRECISION	raw
CHAR	LZO, ZSTD, Text255 and Text32k
VARCHAR	LZO, ZSTD, Text255 and Text32k
DATE	AZ64
TIMESTAMP	AZ64
TIMESTAMPTZ	AZ64
TIME	AZ64

redshift column encoding2.md

encoding	column type
raw	sort keys, BOOLEAN, REAL, DOUBLE PRECISION
AZ64	numeric, date, and time
Byte-dictionary	fewer than 256 unique values
Delta	numeric sequence
LZO	CHAR, VARCHAR, free form text
Mostly	data type for a column is larger than most of the stored values
Runlength	value that is repeated consecutively
Text255 and Text32k	VARCHAR columns in which the same words recur often
ZSTD	ALL, CHAR, VARCHAR