Change Log

2.0.0

  • Added support for like and ilike in the query builder. MySQL already does case insensitive and will not support ilike.
  • Updated schema key listing utility to use built in Ecto __schema__ reflection function. Will now filter out virtual types.
  • Lexical Atom queries have been soft deprecated. There doesn’t seem to be a reason to support this any further as that it does not provide much(if any) benefit over the standard query keyword list or map. This feature will be removed in a future release.
  • Added updated_all and deleted_all functions to store.
  • It has been a while since I last updated the Ecto dependency. It has now been updated to require 2.2 or greater. This has resulted in a major version revision since this will force applications to move up to a newer version of Ecto.

1.9.1

  • Added the ability to add a query to the front of an update statement. This will only update the first record retrieved and not all records.

1.9.0

  • Optimized query builder to utilize dynamic field names in Ecto queries. This will reduce compile time and generated beam size.

1.8.2

  • Order by was overwritting the entire query, fixed expression to build on previous query parameters.

1.8.1

  • Realized that including the module to support rest_api_builder directly in this library creates an unneccessary dependency which also includes references to dependencies used by Phoenix and other frameworks. API provider moved to its own library. rest_api_builder_essp.

1.8.0

  • Adds more advanced event notifications through the EventQueues libary.
  • Adds provider support when using the rest_api_builder library.

1.7.1

  • Added Parser for common changeset errors.

1.7.0

  • Added inserting of struct directly to all insert methods. This will bypass a changeset just as it does in Ecto.
  • Added experimental update_all helper method. Use at your own risk, may not stay around.
  • Update one function to use Ecto.Repo.all and to return the first record instead of throwing an error which Ecto.Repo.one will do. It seems this is a more common use case than wanting an error to be thrown and break your application process.
  • Added order_by options to all and one functions.
  • Renamed destructure to to_map option and function based upon feedback. destructure still exists but points to to_map. Will be removed in a future version.
  • Changed one function receive a String id number as well as an integer. Making it more convient for when an id is passed via web params or read from another text source. No need to convert it before passing it in.

1.6.0

  • Updated to add Ecto 2.1 support with the standard Elixir Date, Time, DateTime, NaiveDateTime structs.
  • Fixed Elixir 1.4.0 warnings.
  • Started changelog.
  • Revised ex_docs with better structure.