Claus Ibsen created CAMEL-26030:
-----------------------------------
Summary: camel-aws2-ddb - wrong header types in the catalog,
CamelAwsDdbNextToken missing, and result headers read by the next step
Key: CAMEL-26030
URL: https://issues.apache.org/jira/browse/CAMEL-26030
Project: Camel
Issue Type: Bug
Components: camel-aws2
Affects Versions: 4.18.0, 4.22.0, 4.23.0
Reporter: Claus Ibsen
Fix For: 4.24.0
Found while adding the catalog body types of aws2-ddb and aws2-ddbstream
(CAMEL-24961). The body is fine: every operation keeps it and puts its result
in headers. The headers are where these are.
h4. 1. The catalog gives wrong types for result headers
The {{javaType}} of these headers in {{Ddb2Constants}} (main 07d2e7a987bc) is
not what the producer sets:
* {{CamelAwsDdbTableStatus}}: String, but the {{TableStatus}} enum of the AWS
SDK (line 131);
* {{CamelAwsDdbCreationDate}}: Date, but a {{java.time.Instant}} (line 50);
* {{CamelAwsDdbConsumedCapacity}}: Double, but a {{ConsumedCapacity}} object,
which is always null, as Query and Scan never ask for it with
ReturnConsumedCapacity (line 44);
* {{CamelAwsDdbLastEvaluatedKey}}: Key, but a {{Map<String, AttributeValue>}}
(line 89);
* {{CamelAwsDdbBatchResponse}}: {{Map<String, BatchResponse>}}, but a
{{Map<String, List<Map<String, AttributeValue>>>}} (line 36);
* {{CamelAwsDdbFilterExpressionAttributeValues}}, an input: {{Map<String,
String>}}, but read as a {{Map<String, AttributeValue>}} (line 205;
{{AbstractDdbCommand}}, line 119).
Checked on 4.24.0-SNAPSHOT with {{camel run}} against the AWS emulator of
{{camel infra run aws}}: DescribeTable gave a TableStatus and an Instant, and
Scan an empty ConsumedCapacity header.
*Proposed change*: correct the javaType of these headers; ask for the consumed
capacity, or drop the header.
h4. 2. CamelAwsDdbNextToken is not in the catalog
ExecuteStatement reads the {{CamelAwsDdbNextToken}} header and sets it to the
token of the next page ({{ExecuteStatementCommand}}, lines 47 and 56), but the
constant has no {{@Metadata}} ({{Ddb2Constants}}, line 74), so the catalog and
the docs do not list it. Checked on 4.24.0-SNAPSHOT with {{camel run}} against
the AWS emulator of {{camel infra run aws}}: ExecuteStatement set it.
*Proposed change*: add the {{@Metadata}}.
h4. 3. The next aws2-ddb step reads headers that a step set as its result
Some result headers are also inputs, so a second aws2-ddb step in a route uses
what the first one gave:
* DescribeTable and DeleteTable set {{CamelAwsDdbTableName}}
({{DescribeTableCommand}}, line 37; {{DeleteTableCommand}}, line 42), which the
next step takes as its table ({{AbstractDdbCommand}}, line 47);
* DescribeTable sets {{CamelAwsDdbReadCapacity}} and
{{CamelAwsDdbWriteCapacity}} (lines 42-43), which a following UpdateTable reads
before its options ({{UpdateTableCommand}}, lines 39-47);
* ExecuteStatement sets {{CamelAwsDdbNextToken}}, which the next
ExecuteStatement sends again, so it reads the next page of another query.
*Proposed change*: remove the input headers a command sets before setting the
result, or name the result headers apart from the inputs.
h4. Also read in the code, not run
* UpdateTable drops the response and sets no header ({{UpdateTableCommand}},
lines 33-36).
* aws2-ddbstream: the section "Sequence Numbers" of the docs describes an
option that CAMEL-16594 removed; the docs say a DynamoDbStreamsClient must be
in the registry, but the access key and the credentials providers work too; the
catalog lists no headers, as {{@UriEndpoint}} of {{Ddb2StreamEndpoint}} has no
headersClass.
* The docs write the uri as {{aws2-ddb://domainName}}; the path is the table
name.
* {{AWS2BatchGetItemsRuleIT}}: the putItem helper sends BatchGetItems, so
nothing is put, and the test only checks that the headers are not null.
_Claude Code on behalf of davsclaus_
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
