View Source Changelog

v0.15.1 (2024-03-18)

• Add feature for local secondary index creation in migration (#44).
• Fix case match to EctoTablestore.Repo.transaction/2 (#45, #46).
• Fix document source link was broken when entry from hexdocs.pm (#47).
• Make macro definition expandable in migration (#48).

v0.15.0 (2024-01-26)

• Add EctoTablestore.Repo.sql_query/2 (#40).
• Add EctoTablestore.Repo.transaction/2 and EctoTablestore.Repo.rollback/1 base on the local transaction (#41, #42).

v0.14.0 (2023-01-04)

• Add Repo.one!/2.
• Enhance Repo.get_range/2, Repo.stream/2 with token option.
• Enhance output when create search index.

v0.13.3 (2022-05-24)

• Fix and update document to Repo's batch get.
• Fix to use the public API Ecto.Adapter.lookup_meta/1 when upgrade to ecto ~> 3.8.*.

v0.13.2 (2022-01-21)

• Fix schema's timestamps() as as the original with second time unit.
• Fix Repo.get_range/4 document missed.

v0.13.1 (2021-10-25)

• Update document to clarify EctoTablestore.Schema use.
• Some changes to schema's timestamps() but with unexpected change the time unit of inserted_at/updated_at fields from second to nanosecond, fixed in 0.13.2, so please DO NOT use this version.

v0.13.0 (2021-08-25)

• Refactor :hashids type with Ecto.ParameterizedType.
• Clean logger when create sequence.
• Fix warning: function checked_out?/1 required with ecto 3.7.0.

v0.12.2 (2021-08-20)

• Some fix and enhancement via #34:
  1. Fix to properly adapt Ecto type on load;
  2. Add support to :decimal type in the field of schema;
  3. Add Ecto.ReplaceableString type.

v0.12.1 (2021-08-11)

• Use schema defined attribute fields into :columns_to_get option as default in all read operations via #33.

v0.12.0 (2021-08-10)

• Some fix and enhancement via #32:
  1. Fix fail when batch write with embeds_many or embeds_one fields;
  2. Add :entity_full_match as an optional option to leverage the provided attribute columns of the struct into the filter of the batch write condition when it is true, default value is false. After this breaking change, each put/update/delete operation of batch write requires explicitly set :condition option, excepts the following #3 case;
  3. When insert or a :put operation of batch write with a table defined an auto increment primary key(a server side function), we can omit :condition option, because the server side only accepts condition: condition(:ignore), we internally wrapper this to simplify the input in this use case;
  4. Clarify document about the :condition option of insert/update/delete/batch_write;
  5. Some code naming improvement.

v0.11.2 (2021-05-21)

• Add drop_if_exists/1 function for migration.
• Add --migrations_path option to mix ecto.ots.migrate.

v0.11.1 (2021-05-17)

• Fix format error to put the .formatter.exs file into package.

v0.11.0 (2021-05-14)

• [Need to migrate before upgrade] When define the type of partition key as an auto increment integer, we created a "#{table_name}_seq" table as an internal sequence to the "#{table_name}" table, but now all sequence(s) managed into the "ecto_tablestore_default_seq" table, see #16.
• Add Repo.stream_search/3, see #17.
• Add Repo.stream/2, see #21.
• Add Repo.get_range/1 and Repo.get_range/2 for easy use, see #28.
• Support changeset optimistic_lock in update, see commit 66f6823.
• [Need to check before upgrade] Fix update with :increment operation return an unexpected result when schema has multi integer fields defined, and require to explicitly add column name(s) which are related in the atomic increment operation into the :returning option of Repo.update/2, see #19.
• Fix to properly handle the stale error case when insert, see #20.
• Fix to remove the useless supervisor name to resolve duplicated naming conflict in some cases, see #24.