- Add python 3.10 compatibility
- Bump
joblib
from 0.16.0 to 1.2.0 - Bump
psycopg2-binary
from 2.8.5 to 2.9.5 to get macOS arm64 (Apple M1) wheels
- Bump
joblib
from 0.13.2 to 0.16.0 to be Python 3.8 compatible
- Enable SSL mode if
ssl
option is"true"
in config
- Switch jsonschema to use Draft7Validator
WARNING: This release includes non backward compatible changes.
Starting from pipelinewise-target-postgres-2.0.0
the integer
JSON Schema column types with minimum and maximum
boundaries are loaded into Postgres SMALLINT
, INTEGER
and BIGINT
values. If you're upgrading from an
earlier version of pipelinewise-target-postgres then it's recommended to re-sync every table otherwise all the existing
NUMERIC
columns in Postgres will be versioned to the corresponding integer type.
Further info about versioning columns at https://transferwise.github.io/pipelinewise/user_guide/schema_changes.html?highlight=versioning#versioning-columns
- Add
flush_all_streams
option - Add
parallelism
option - Add
max_parallelism
option - Add
validate_records
option - Log inserts, updates and csv size_bytes in a more consumable format
- Fixed an issue when JSON values sometimes not sent correctly
- Support usage of reserved words as table and column names
- Add
temp_dir
optional parameter to config - Load
integer
JSON Schema types with min and max boundaries to PostgresSMALLINT
,INTEGER
,BIGINT
column types - Switch to
psychopg-binary
2.8.5
- Support custom logging configuration by setting
LOGGING_CONF_FILE
env variable to the absolute path of a .conf file
- Grant privileges correctly when table created
- Fixed type mapping
- Add license details
- Sync column versioning feature to other supported targets
- Sync data flattening option to other supported targets
- Sync metadata columns behaviour to other supported targets
- Bump
psycopg2
to 2.8.2
- Initial release