Release Note 20151101
This is a summary of new features and improvements introduced in the November 1, 2015 release:
Table of Contents
Collection: Data Connector for Mixpanel
By using this feature, you can import data from Mixpanel without coding anything. You can schedule daily data import as well.
For further information visit Data Connector for Mixpanel.
Output: Result Output to BigQuery
Output: Result Output to Tableau for v9.1
re:dash is an open-source web application that allows to easily query an existing database, share the dataset and visualize it in different ways, and create dashboards. From re:dash v8.1, it supports Treasure Data as one of its data sources.
Hive: TD_DIVIDE UDF
We have added the TD_DIVIDE UDF to the Hive engine, which makes the division of numeric types safer even when the denominator is zero. If the denominator is zero, it returns 0; in all other cases it returns the result of the division between the 2 arguments as expected.
See the documentation page for more info: TD_DIVIDE(num, den)
Hive: TD_SUBSTRING_INENCODING UDF
We have added the TD_SUBSTRING_INENCODING UDF to the Hive engine; it returns the substring or slice of the byte array in input from the 0-index position at most max_len_inbytes with charset encoding. charset is one of those indicated here: java.nio.charset.Charset.
See the documentation page for more info: TD_SUBSTRING_INENCODING(column, length, charset)
Machine Learning: Random Forest & Factorization Machine
Hivemall was upgraded to v0.4.0-1, from v0.3.2-3. This major release supports two major algorithm: Random Forest & Factorization Machine.
- Doc: Iris multi class classification using RandomForest
- Doc: Movielens Rating Prediction using Factorization Machine
Client: td-client-node v0.2.3
REST API client for node.js v0.2.3 was released.
See the package repository here: td v0.2.3
Last modified: Nov 16 2015 12:32:27 UTC