- Update document to clarify
EctoTablestore.Schema
use.
- Refactor
:hashids
type with Ecto.ParameterizedType. - Clean logger when create sequence.
- Fix warning: function checked_out?/1 required with ecto
3.7.0
.
-
Some fix and enhancement via #34:
- Fix to properly adapt Ecto type on load;
- Add support to
:decimal
type in the field of schema; - Add
Ecto.ReplaceableString
type.
- Use schema defined attribute fields into
:columns_to_get
option as default in all read operations via #33.
-
Some fix and enhancement via #32:
- Fix fail when batch write with
embeds_many
orembeds_one
fields; - 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 istrue
, default value isfalse
. After this breaking change, each put/update/delete operation of batch write requires explicitly set:condition
option, excepts the following#3
case; - 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 acceptscondition: condition(:ignore)
, we internally wrapper this to simplify the input in this use case; - Clarify document about the
:condition
option of insert/update/delete/batch_write; - Some code naming improvement.
- Fix fail when batch write with
- Add
drop_if_exists/1
function for migration. - Add
--migrations_path
option tomix ecto.ots.migrate
.
- Fix format error to put the
.formatter.exs
file into package.
- [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
andRepo.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 ofRepo.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.