Release Note 20140617

Table of Contents

Features & Improvements

This is a summary of the new features and improvements introduced in this release:

Console: Android SDK

Added the Android SDK icon to the Import Data, SDKs landing page to introduce the first release (v0.1.0) of the SDK.

Console: Android SDK

Console: Filter Jobs By Executing User

Added the ability to filter the list of Jobs in the Jobs list page by any Executing User.

Previously the filtering was limited to the current logged on user only.

Console: Filter Jobs By Executing User

Console: Delete Saved Query Result Output Target

Added the ability to delete a previously saved Named Query Result Output Target. The user can delete the target from within the New Query page.

Console: Delete Saved Query Result Output Target

SDKs: Released Android SDK v0.1.0

The first official version of the Android SDK has been released this week as v0.1.0.
The release was also published in the Central Maven repository on Sonatype.

For more information please visit the Android SDK documentation page.

SDKs: Released iOS SDK v0.0.6

This is an incremental release of the iOS SDK (first released last week) with some enhancements:

  • Support compression of the body of HTTP requests;
  • Added the SDK version in the HTTP User-Agent;
  • Added support for importing each chunk of data with an unique event ID associated to it;
  • Added validation of the Treasure Data database and table name formats;
  • Added support for importing with multiple account IDs from within the same application.

Bulk Import: Released v0.4.14

See Bug Fixes, Bulk Import section.

Toolbelt: Released Ruby CLI v0.11.2

  • Added a work around for a problem in the Pig Engine which causes it to return a ‘null’ column name for column values generated by Pig functions (e.g. SUM, COUNT, …);
  • Removed Query Result Output Target to Treasure Data database and table name format validation: it’s already done by the APIs;
  • Added the ability to specify an alternative API endpoint in the configuration file and at runtime with the -e / --endpoint options.

See also the Bug Fixes, Toolbelt section.


Bug Fixes

These are the most important Bug Fixes made in this release:

APIs: User Creation Endpoint

  • [Problem]
    The ‘user/add/:name’ REST API to invite users to an existing Treasure Data account returns 500 errors (Internal Server Error) errors when called with any REST API client.
    Account owners and administrators are still able to invite users through the Console interface.
    [Solution]
    The expected format of the payload from the client had changed, breaking backwards compatibility. The clients had not been updated accordingly.
    Restored the ‘/users/add/:name’ API backwards compatibility. No change to the client libraries was necessary.

Console: File Uploader Support for Firefox and IE

  • [Problem]
    The Import Data, File uploader does not work reliably on Firefox and Internet Explorer. When the finalized import is submitted for processing, an error concerning the imported file’s name and size is displayed.
    [Solution]
    This is caused by submitting the import request without file name and size information.
    File name and size is now submitted along with the rest of the information. The uploader was tested on both Internet Explorer (IE11 on Windows 8.1) and Firefox (v30 on MacOSX 10.8.5).

Console: Pig Engine Reserved Keywords

  • [Problem]
    Pig Engine returns a “Syntax error, unexpected symbol at or near ‘=’
    [Solution]
    This error occurs when a table is named after one of the reserved keywords in Pig (‘sample’ for example).
    Pig was modified to ignore column names which match reserved keywords in the queried tables' schema.

Bulk Import: Fail If No Source

  • [Problem]
    Import attempts do not fail if no source is provided.
    [Solution]
    There is no validation on whether the set of options the Bulk Import tool is called with contain at least a valid file name, MySQL table, or S3 path.
    The check has been added and the import now fails if no valid source is specified.

Bulk Import: Import ‘auto’ From AWS S3

  • [Problem]
    Import import:auto cannot import data from AWS S3.
    [Solution]
    The import:auto command (unlike import:prepare which works since it does not require interaction with the Treasure Data REST APIs) fails authenticating with the AWS S3 public/private keys embedded in the source path.
    This is due to a conflict between the Treasure Data and AWS S3 client side certificates.
    The certificate verification for AWS S3 was disabled to allow the communication to take place.

Toolbelt: Proxy Issues

  • [Problem]
    Bulk Import JAR file and Toolbelt update don’t honor the HTTP_PROXY environment variable setting.
    [Solution]
    The HTTP_PROXY environment variable setting is now taken into consideration when setting up the HTTP communication with the update endpoints for the JAR file and Toolbelt.

  • [Problem]
    Bulk Import commands are not executed through a proxy if an HTTP_PROXY is specified.
    [Solution]
    The HTTP_PROXY environment variable setting is used to setup the HTTP communication with the Bulk Import REST API endpoints from within the Java Client Library.

Toolbelt: Bulk Import Version Update

  • [Problem]
    When the Bulk Import JAR file update fails, the Bulk Import tool is no longer usable.
    [Solution]
    When the Bulk Import JAR file updater fails because of a network connectivity error, no exception is raised and the flow continues in spite of the fact no updated Bulk Import binary can be retrieved. Ultimately, the existing Bulk Import JAR file is overwritten with an invalid binary, rendering the tool unusable.
    The updater now checks the status after each attempt to fetch data from the update endpoint. In case of failure, the Bulk Import JAR file is rolled back to the previous version.

Last modified: Feb 22 2017 23:40:42 UTC

If this article is incorrect or outdated, or omits critical information, please let us know. For all other issues, please see our support channels.