Skip to content

Tags: rishindoshi/PynamoDB

Tags

5.3.3

Toggle 5.3.3's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Improve first iteration logic (pynamodb#1121)

Backporting pynamodb#1101 to 5.x branch to allow handling an exception raised when retrieving the first item.

5.3.2

Toggle 5.3.2's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Do not package typing_tests (pynamodb#1118)

5.3.1

Toggle 5.3.1's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Upgrade to mypy 0.950 (pynamodb#1116)

mypy<0.900 is dragging in typed-ast which no longer seems to build on Python 3.11. In order to unblock the release of 5.3.1, we'll pull up to mypy==0.950 (for assert_type) and drag in some minor (non-breaking) improvements from master.

5.3.0

Toggle 5.3.0's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Do not call DescribeTable for models (pynamodb#1095)

Do not call DescribeTable for models (pynamodb#1091)

When using DynamoDB through a `Model` or `Index` (rather than `(Table)Connection` directly), we will derive the "meta-table" from the model itself rather than make an initial `DescribeTable` call. This has numerous advantages:
- Faster bootstrap (important for lambdas, as pointed out in pynamodb#422)
- More consistent handling of attribute types: Before this change, if the PynamoDB model definition and the DynamoDB table definition disagreed on a key attribute's type, PynamoDB would use its own idea of the type in some code paths and the underlying type in others. Now it would consistently use its own idea of the type, allowing the erroneous model definition to be spotted sooner.
- Easier testing, since there's no longer a one-off request that only happens once and affects global state.

This approach attempts to change the library as little as possible, by synthesizing a MetaTable from the model.

This is a backport of pynamodb#1091.

5.2.3

Toggle 5.2.3's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Workaround for botocore 1.28 (round 2) (pynamodb#1087)

In pynamodb#1083 we've started passing an `endpoint_url` parameter to _convert_to_request_dict due to changes made in botocore 1.28.

When a model does not specify a `host`, the `endpoint_url` would be `None`. To determine the actual `endpoint_url` in botocore ≥1.28, we must call another private method, `_resolve_endpoint_ruleset`.

5.2.2

Toggle 5.2.2's commit message
mention the _convert_to_request_dict exception

5.2.1

Toggle 5.2.1's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Prepare for 5.2.1 release (pynamodb#1019)

5.2.0

Toggle 5.2.0's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Prepare for 5.2.0 release (pynamodb#1010)

5.1.0

Toggle 5.1.0's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Release Notes : 5.1.0 (pynamodb#956)

* 5.1.0

Co-authored-by: Garrett <[email protected]>

5.0.3

Toggle 5.0.3's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
Propagate null_check flag when serializing nested MapAttributes (pyna…

…modb#908)

Before this change, the Model would propagate null_check to the immediate MapAttributes but then those MapAttributes wouldn't propagate it to deeper-nested ones.