This bug is reported by @maartenpants on gitter. Original report is here:
Hi there, has anyone else experienced this issue with CmaEsSampler and the PostgreSQL storage backend:
(psycopg2.errors.StringDataRightTruncation) value too long for type character varying(2048)
[SQL: INSERT INTO trial_system_attributes (trial_id, key, value_json) VALUES (%(trial_id)s, %(key)s, %(value_json)s) RETURNING trial_system_attributes.trial_system_attribute_id]
It seems the serialized version of the CMA class is too large to fit in the 2048 character column. This code is the culprit:
This result means that the bug is raised even on low-dimensional search space.
Additional context (optional)
I've ever tested on SQLite3 but never bumped into this bug. The reason why is:
(9) What is the maximum size of a VARCHAR in SQLite?
SQLite does not enforce the length of a VARCHAR. You can declare a VARCHAR(10) and SQLite will be happy to store a 500-million character string there. And it will keep all 500-million characters intact. Your content is never truncated. SQLite understands the column type of "VARCHAR(N)" to be the same as "TEXT", regardless of the value of N.
https://www.sqlite.org/faq.html#:~:text=(9)%20What%20is%20the%20maximum,all%20500%2Dmillion%20characters%20intact.
Problem details
This bug is reported by @maartenpants on gitter. Original report is here:
https://gitter.im/optuna/optuna?at=5f484dea36e6f709fd08a5fa
I checked the length of serialized CMA object.
This result means that the bug is raised even on low-dimensional search space.
Additional context (optional)
I've ever tested on SQLite3 but never bumped into this bug. The reason why is: