Skip to content

Latest commit

 

History

History
73 lines (68 loc) · 6.3 KB

CHANGELOG.md

File metadata and controls

73 lines (68 loc) · 6.3 KB

Changelog

[3.8.0] - 2024-10-09

Welcome to the MCQA 2.0 Era!

New Features

  • Support for submitting MCQA 2.0 jobs. Enable MCQA 2.0 mode by setting quotaName to interactive quota and turning on interactiveMode=true.
  • New parameter useInstanceTunnel, which specifies whether to use Tunnel to retrieve data. The default value is true, consistent with previous behavior. If instanceTunnel is turned off, the system will use RESTful API to retrieve data under this mode:
    • Execution speed will be faster.
    • Only 10,000 results can be returned; if this limit is exceeded, the data will be truncated.
    • Return values are of String type, suitable only for display-type jobs.
  • New executable class JdbcRunner, which allows executing JDBC tasks via a connection string and SQL file. The usage is as follows:
 java -cp odps-jdbc-version-jar-with-dependencies.jar com.aliyun.odps.jdbc.JdbcRunner <jdbc_url> <sql_file>
  • Added some debug settings that modify the current Connection configuration (effectively re-acquiring the Connection to replace the current one).
    • set tunnelEndpoint = xxx; modifies the tunnelEndpoint parameter of the current Connection.
    • set useTunnel = true/false; modifies the useInstanceTunnel parameter of the current Connection.
    • set interactiveMode = true/false; modifies the interactiveMode parameter of the current Connection.

Changes

  • Removed handling of # comment in the removeComment method: it no longer removes comments after the #. This change resolves many issues where valid SQL statements were incorrect due to the removal of #.
  • Updated odps-sdk version to 0.50.0-public.

[3.7.0] - 2024-08-29

Compatibility with Metabase!
Learn how to connect MaxCompute to Metabase

Important Changes

  • Optimized the DatabaseMetadata.getColumns Logic: Enhanced compatibility for three-tier and two-tier model users as follows:
    • Three-tier model (project.schema.table):
      • Ignores the passed catalog, using the project name from the JDBC connection, along with the schemaPattern and tableNamePattern parameters to compose the table reference.
    • Two-tier model (project.table):
      • Ignores the schemaPattern, treating the catalog as the project name, and combines it with the tableNamePattern parameter to form the table reference.
    Three-tier model: catalog.schemaPattern.tableNamePattern -> project.schemaPattern.tableNamePattern
Two-tier model: schemaPattern.tableNamePattern -> catalog.tableNamePattern

Changes

  • Enhanced SQLException Messages: The exception message now includes the correct ErrorMessage.
  • Updated odps-sdk Version: Updated to 0.48.8-public.

[3.6.0] - 2024-08-07

Say No to SQLFeatureNotSupportedException!

New Features

  • LocalDate Support: PrepareStatement.setObject now supports LocalDate type for MaxCompute DATE type.
  • SQL Injection Check: New parameter skipSqlInjectCheck allows skipping the SQL injection check, applicable when setting String type fields.
  • TIMESTAMP_NTZ Support: DatabaseMetadata now also supports MaxCompute TIMESTAMP_NTZ type fields.
  • Extended PrepareStatement Methods:
    • Implemented setObject(int, Object, int), setObject(int, Object, int, int), setObject(int, Object, SQLType), and setObject(int, Object, SQLType, int) methods, which will no longer throw exceptions but will return results from setObject(int, Object), ignoring additional fields.
  • Extended ResultSet Methods:
    • Added getObject(int, Map) and getObject(String, Map) methods, which no longer throw exceptions but ignore the map and call ResultSet.getObject(int) to return results.
    • Added getObject(int, Class<T>) and getObject(String, Class<T>) methods, which first call ResultSet.getObject(int) before trying to convert the result to the specified type.
  • Optimized Statement Method: The createStatement(int, int, int) method now no longer throws exceptions but ignores resultSetHoldability, returning the result of Statement.createStatement(int, int).

Changes

  • Enhanced DatabaseMetadata.getColumns: The interface now returns IS_AUTOINCREMENT and IS_GENERATEDCOLUMN fields to comply with JDBC specifications.
  • Updated odps-sdk Version: Updated to 0.48.7-public.
    • The new version applies key-path-end optimizations, improving efficiency during offline job execution, especially in complex queries.

[3.5.8] - 2024-07-22

New Features

  • Delta Table Write Support: PrepareStatement now supports write operations on Delta Tables.
  • Skip SQL Rewrite: New parameter skipSqlRewrite, with a default value of true, allows skipping SQL statement rewriting during comment removal. Enabling this parameter may cause settings submitted with the SQL to become ineffective.

Changes

  • Optimized Comment Removal Logic: The logic has been optimized to avoid throwing exceptions when processing very long queries. It fixes the issue of potentially removing SQL hints during comment removal.
  • SDK Update: The odps-sdk has been updated to 0.48.6-public.
    • The new version reduces network overhead, slightly improving efficiency during offline job execution.

[3.5.7] - 2024-04-29

New Features

  • New Parameter tunnelDownloadUseSingleReader: The default value is false. When this parameter is enabled, each ResultSet will use only a single Reader for data downloading. This change is suitable for scenarios that require reading large amounts of data at once, optimizing performance. The previous behavior was to open a new Reader each time hasNext is called, to prevent connection drops.

[3.5.6] - 2023-10-27

New Features

  • Support for insert into tablename (co1, co2) values(?, ?) syntax.
  • DatabaseMetadata.getColumns now adapts to situations using the three-tier model.