compression difference
Navin_C
Vertica Customer ✭
Hello All,
This is a part of some test performed.
We tried replicating the data in same cluster with different schema name.
We have more then 10 schema for doing this
So for schema A having 30 tables
we replicated schema B having same 30 tables in same cluster
How we did this:
1. export the definition of 30 tables along with all projection definition
2 Changed the definition(name) and deployed the ddl into same cluster
3. inserted all records from schema A to schema B table wise using insert..select statement
According to these steps all the things should be same in both schemas
Differeces observed
1. One of the schema (B) is doubled in terms of compressed size from Schema A
2. Others schemas (B) are reduced in terms of compressed size in range of 4-5 GB from Schemas A
We have checked the rowcount for each tables and they are perfectly matching their corresponding tables in other schema
Got the compressed size from projection storage table
Why is there such a difference in compression size.
This is a part of some test performed.
We tried replicating the data in same cluster with different schema name.
We have more then 10 schema for doing this
So for schema A having 30 tables
we replicated schema B having same 30 tables in same cluster
How we did this:
1. export the definition of 30 tables along with all projection definition
2 Changed the definition(name) and deployed the ddl into same cluster
3. inserted all records from schema A to schema B table wise using insert..select statement
According to these steps all the things should be same in both schemas
Differeces observed
1. One of the schema (B) is doubled in terms of compressed size from Schema A
2. Others schemas (B) are reduced in terms of compressed size in range of 4-5 GB from Schemas A
We have checked the rowcount for each tables and they are perfectly matching their corresponding tables in other schema
Got the compressed size from projection storage table
Why is there such a difference in compression size.
0
Comments
If you assign an ENCODING value in the CREATE PROJECTION statement, that ENCODING value gets exported properly. If an ENCODING value is not specified, then AUTO is assumed, which can lead to the issue you have specified. The database designer typically lists an ENCODING value for fields in a projection.
Please check "projection_columns" and "column_storage" tables for verifying if encoding is not changed in new schema.
Regards'
Abhishek
As I said about the steps.
I have exported the actual projection definition and then deployed it.
So whatever the actual encoding is for Schema A will be same for Schema B.
Thanks