All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Add unicode_string_param function #60
- ODBC error not always a string #53
- Handle UTF-16 return data from Snowflake #38
- General improvements, add DBConnection adapter #32
- Callback that calls gc after queries to try to shuffle binary strings off the heap. We've seen large memory savings in a real application. #23
- Relax dependency requirement on
telemetry
- Hot fix for possible memory leak #27
- Properly convert elixir nil to erlang :null in query params
- Added telemetry 1.0
- New telemetry events around queries
- Enabled SQL
INSERT
statements.
- Corrected spelling in type definition
- Added
map_nulls_to_nil?
variable to connection configuration to allow conversion of:null
values to:nil
in snowflake query response
- Initialize the worker with a propslist instead of a tuple.
- create copies of the
odbc
type definitions
- Added the ability to keep connections alive through configuration on the
Snowflex.Connection
module.
- make sure to follow
{:error, reason}
convention in all parts of the worker
- fix bug where handle_info is missing argument
- Added the ability to define a module that will use
Snowflex.Connection
to maintain connection information
- Remove
Snowflex.Query
as it was a duplicate - Remove
Snowflex.ConnectionPool
in favor of newSnowflex.Connection
- Configure the worker module on the
Snowflex.ConnectionPool
and no longer on the application level.
- Users must launch connection pools as part of their application's
supervision tree using
Snowflex.ConnectionPool
, and all queries must specify which connection pool to use. No connection pool will be started by default.