-
Notifications
You must be signed in to change notification settings - Fork 15
Description
Currently for redshift and snowflake the domain_sessionid column is loaded as a char(128) when most other columns are a varchar(128). In snowflake this doesn't actually matter as snowflake does not pad the string (https://docs.snowflake.com/en/sql-reference/data-types-text#char-character-nchar) but in redshift this padding does occur, which can lead to confusion from users and edge case issues with data models, it also has a slightly higher storage cost when compared to a varchar.
Given no other columns are a char I don't foresee a reason that this needs to remain as one.
This is everywhere I can see the definition made:
Line 134 in 05dd5ab
| | domain_sessionid char(128) encode ZSTD, |
Line 134 in 05dd5ab
| | domain_sessionid CHAR(128), |
| domain_sessionid CHAR(128), |
Line 193 in 05dd5ab
| Column("domain_sessionid", Char(128)), |