You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@camel.apache.org by "Claus Ibsen (Jira)" <ji...@apache.org> on 2026/10/23 05:20:00 UTC

[jira] [Created] (CAMEL-26056) camel-jdbc - an outputClass that does not resolve fails with a NullPointerException, and generated keys use outputClass

Claus Ibsen created CAMEL-26056:
-----------------------------------

             Summary: camel-jdbc - an outputClass that does not resolve fails with a NullPointerException, and generated keys use outputClass
                 Key: CAMEL-26056
                 URL: https://issues.apache.org/jira/browse/CAMEL-26056
             Project: Camel
          Issue Type: Bug
          Components: camel-jdbc
    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 jdbc (CAMEL-24961), checked with {{camel run}} with an H2 in-memory data source.

h4. 1. An outputClass that does not resolve fails with a NullPointerException

{{JdbcHelper}} resolves the class of outputClass and creates it without a check ({{JdbcHelper}}, main 5aac7df76b83, lines 35-36), so a name that does not resolve fails with {{RuntimeCamelException: NullPointerException: Cannot invoke "java.lang.Class.getDeclaredConstructor(...)" because "type" is null}}, without the class name. Checked on 4.24.0-SNAPSHOT with {{camel run}}: outputClass=survey.NoSuchItem.

*Proposed change*: fail with a message that names the class that is not found.

h4. 2. Generated keys are mapped to outputClass

The generated keys go through {{extractRows}} (JdbcProducer, line 375), which applies outputClass and readSize (line 416 on). So with outputClass on the endpoint, {{CamelGeneratedKeysRows}} holds beans of that class, or the insert fails with "Cannot map all properties" when the key columns are not its properties, and with resetAutoCommit the insert is rolled back (read in the code). Checked on 4.24.0-SNAPSHOT with {{camel run}}: an insert with CamelRetrieveGeneratedKeys on an endpoint with outputClass gave beans of that class with only the id set.

h4. 3. A Blob cannot be read after the producer

A BLOB or CLOB value is a {{java.sql.Blob}} or {{Clob}} ({{ResultSetIterator}}, line 213), but the producer closes the connection before it returns, so the route cannot read it. useGetBytesForBlob gives a byte[] for a BLOB; nothing does it for a CLOB. Checked on 4.24.0-SNAPSHOT with {{camel run}}: a SelectOne of a BLOB column gave an {{org.h2.jdbc.JdbcBlob}}, and its length then failed with "The object is already closed".

h4. 4. A null body fails with a NullPointerException

The SQL is read with {{getBody(String.class)}} (lines 77 and 125), not mandatory, so no body fails with useHeadersAsParameters with {{NullPointerException: Cannot invoke "java.lang.CharSequence.length()" because "this.text" is null}}, and otherwise with the error of the driver. Checked on 4.24.0-SNAPSHOT with {{camel run}}.

*Proposed change*: read the body with {{getMandatoryBody(String.class)}}.

h4. 5. The docs say generated keys do not work with named parameters

{{jdbc-component.adoc}} line 75: "Using generated keys does not work together with named parameters.". They do: Checked on 4.24.0-SNAPSHOT with {{camel run}}, an insert with {{:?label}} and useHeadersAsParameters gave its key in CamelGeneratedKeysRows, and JdbcParameterizedQueryGeneratedKeysTest tests it.

h4. Also read in the code, not run

* The transacted option is not read anywhere ({{JdbcEndpoint}}, lines 49 and 116-123), since CAMEL-3803.
* CamelJdbcRowCount is set for SelectList only, but its doc says for a select.
* Only the first result of a statement is read (no getMoreResults).
* StreamList with resetAutoCommit commits before the route reads the rows, which a driver that closes cursors at commit cannot read.

_Claude Code on behalf of davsclaus_



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)