Engineering Data Analytics with Presto and Parquet at Uber
Data / MLEngineering Data Analytics with Presto and Apache Parquet at UberJuly 11, 2017 / Global From determining the most convenient rider pickup points to predicting the fastest routes, Uber uses data-driven analytics to create seamless trip experiences. Within engineering, analytics inform decision-making processes across the board. As we expand to new markets, the ability to accurately and qui
Hadoop Query Performance Smackdown
The document presents a comprehensive performance evaluation of various SQL query engines within a big data environment at Comcast, detailing the test setup and methodology used to assess their efficiency with TPC-DS datasets. Key findings reveal that LLAP exhibited the fastest execution times, outperforming Presto and Tez, while MapReduce and the Spark Thrift Server were identified as underperfor
Python + Hive on AWS EMR で貧者のログサマリ
Sep 14, 201422 likes6,707 viewsAI-enhanced description 1. Akira Chiku is an engineer who works on an engineering team. Their requirements include collecting between 10-20GB of data per day from various sources like Hadoop and Hive. 2. Data is collected from sources like Fluentd and parsed using Query String and stored in Hive. It is then processed and visualized. 3. Data can be stored in S3, proce
MPP on Hadoop, Redshift, BigQuery - Go ahead!
Twitterで「早く今流行のMPPの大まかな使い方の違い書けよ！」というプレッシャーが半端ないのでてきとうに書きます．この記事は俺の経験と勉強会などでユーザから聞いた話をもとに書いているので，すべてが俺の経験ではありません(特にBigQuery)．各社のSAの人とかに聞けば，もっと良いアプローチとか詳細を教えてくれるかもしれません． オンプレミスの商用MPPは使ったことないのでノーコメントです． MPP on HadoopでPrestoがメインなのは今一番使っているからで，Impalaなど他のMPP on Hadoop的なものも似たような感じかなと思っています． もちろん実装の違いなどがあるので，その辺は適宜自分で補間してください． 前提 アプリケーションを開発していて，そのための解析基盤を一から作る． 簡単なまとめ データを貯める所が作れるのであれば，そこに直接クエリを投げられるPre
What are the main differences between Facebook, Presto, and Amplab Shark?
Answer (1 of 2): 1. Primary Use Case: While both are intended for analytics, Shark's primary use case is providing SQL to an (extremely fast) in-memory database, with support also for on-disk (or abstract) data sources. Presto is designed to be a fast SQL engine for the latter, and does not have ...
1
リリース、障害情報などのサービスのお知らせ
最新の人気エントリーの配信
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
Dynamic filtering for highly-selective join optimization
GitHub - Netflix/iceberg: Iceberg is a table format for large, slow-moving tabular data
Prestoで実現するインタラクティブクエリ - dbtech showcase 2014 Tokyo
Prestoソースコードリーディング