The document discusses Apache HCatalog, which provides metadata services and a unified view of data across Hadoop tools like Hive, Pig, and MapReduce. It allows sharing of data and metadata between tools and external systems through a consistent schema. HCatalog simplifies data management by allowing tools to access metadata like the schema, location, and format of data from a shared metastore instead of encoding that information within each application.

1. Toronto Hadoop User Group
February 20, 2013
Apache HCatalog Overview
& Next-gen Hive
Adam Muise
© Hortonworks Inc. 2012 Page 1

2. HCatalog
© Hortonworks Inc. 2012 Page 2

3. Apache HCatalog
• Incubator Project at Apache.org
• Good adoption
• Will likely merge with Hive project as it adds
important functionality to metastore
• Allows for a “schema-on-read” approach to Big Data
in HDFS
• Seat of a lot of innovation in Data Management
• Used by platform partners to enhance Hadoop
integration
• Will likely be used to enhance existing Data
Management products in the Enterprise & create new
products
© Hortonworks Inc. 2012 Page 3

4. Great Tooling Options
MapReduce
• Early adopters
• Non-relational algorithms
• Performance sensitive applications
Pig
• ETL
• Data modeling
• Iterative algorithms
Hive
• Analysis
• Connectors to BI tools
Strength: Right tool for right application
Weakness: Hard to share their data
© Hortonworks Inc. 2012 Page 4

5. HCatalog Changes the Game
Apache HCatalog provides flexible metadata
services across tools and external access
• Consistency of metadata and data models across the
Enterprise (MapReduce, Pig, Hbase, Hive, External Systems)
• Accessibility: share data as tables in and out of HDFS
• Availability: enables flexible, thin-client access via REST API
HCatalog Shared table
and schema
management
• Raw Hadoop data Table access opens the
• Inconsistent, unknown Aligned metadata platform
• Tool specific access REST API
© Hortonworks Inc. 2012 Page 5

6. Options == Complexity
Feature MapReduce Pig Hive
Record format Key value pairs Tuple Record
Data model User defined int, float, string, int, float, string,
bytes, maps, maps, structs, lists
tuples, bags
Schema Encoded in app Declared in script Read from
or read by loader metadata
Data location Encoded in app Declared in script Read from
metadata
Data format Encoded in app Declared in script Read from
metadata
• Pig and MR users need to know a lot to write their apps
• When data schema, location, or format change Pig and MR apps must be
rewritten, retested, and redeployed
• Hive users have to load data from Pig/MR users to have access to it
© Hortonworks 2012
Page 6

7. Hcatalog == Simple, Consistent
Feature MapReduce + Pig + HCatalog Hive
HCatalog
Record format Record Tuple Record
Data model int, float, string, int, float, string, int, float, string,
maps, structs, lists bytes, maps, maps, structs, lists
tuples, bags
Schema Read from Read from Read from
metadata metadata metadata
Data location Read from Read from Read from
metadata metadata metadata
Data format Read from Read from Read from
metadata metadata metadata
• Pig/MR users can read schema from metadata
• Pig/MR users are insulated from schema, location, and format changes
• All users have access to other users’ data as soon as it is committed
© Hortonworks 2012
Page 7

8. Hadoop Ecosystem
Hive Pig MapReduce
(SQL) (scripting) (Java)
Interface: Interface: Interface:
SQL SerDe Load/Store
DML
Input/ Input/ Input/
OutputFormat OutputFormat OutputFormat
DDL
metastore dn1 dn2 dn3 . . .
- tables
- partitions
- files . . . . . .
- types
. . . . dnN
HDFS
© Hortonworks Inc. 2012

9. Opening up Metadata to MR & Pig
Hive Pig MapReduce
(SQL) (scripting) (Java)
HCat Metadata layer
Interface: Interface:
SQL HCatLoad/Store
Interface:
SerDe
HCatInput/OutputFormat
metastore dn1 dn2 dn3 . . .
- tables
- partitions
- files . . . . . .
- types
. . . . dnN
HDFS
© Hortonworks Inc. 2012

10. Pig Example
Assume you want to count how many time each of your users went to each of your URLs
raw = load '/data/rawevents/20120530' as (url, user);
botless = filter raw by myudfs.NotABot(user);
grpd = group botless by (url, user);
cntd = foreach grpd generate flatten(url, user), COUNT(botless);
No need to know No need to
store cntd into '/data/counted/20120530';
file location declare schema
Using HCatalog:
raw = load 'rawevents' using HCatLoader();
botless = filter raw by myudfs.NotABot(user) and ds == '20120530';
grpd = group botless by (url, user);
cntd = foreach grpd generate flatten(url, user), COUNT(botless);
store cntd into 'counted' using HCatStorer();
Partition filter
© Hortonworks 2012
Page 10

11. Working with HCatalog in MapReduce
Setting input: table to read from
HCatInputFormat.setInput(job,
InputJobInfo.create(dbname, tableName, filter));
database to read specify which
Setting output:
from partitions to read
HCatOutputFormat.setOutput(job,
OutputJobInfo.create(dbname, tableName, partitionSpec));
specify which
Obtaining schema: partition to write
schema = HCatInputFormat.getOutputSchema();
access fields by
Key is unused, Value is HCatRecord: name
String url = value.get("url", schema);
output.set("cnt", schema, cnt);
© Hortonworks 2012
Page 11

12. Managing Metadata
• If you are a Hive user, you can use your Hive metastore with no modifications
• If not, you can use the HCatalog command line tool to issue Hive DDL (Data
Definition Language) commands:
> /usr/bin/hcat -e ”create table rawevents (url string,
user string) partitioned by (ds string);";
• Starting in Pig 0.11, you will be able to issue DDL commands from Pig
© Hortonworks 2012
Page 12

13. Templeton - REST API
• REST endpoints: databases, tables, partitions, columns, table properties
• PUT to create/update, GET to list or describe, DELETE to drop
Get a list of all tables in the default database:
Create new table “rawevents”
Hadoop/
HCatalog
Describe table “rawevents”
© Hortonworks 2012
Page 13

14. HCatalog is in Hortonworks Data Platform…
OPERATIONAL
SERVICES
DATA
SERVICES
AMBARI
FLUME
PIG
HIVE
HBASE
SQOOP
HCATALOG
OOZIE
WEBHDFS
MAP
REDUCE
HADOOP
CORE
HDFS
YARN
(in
2.0)
Enterprise Readiness
PLATFORM
SERVICES
High Availability, Disaster Recovery, Snapshots, Security, etc…
HORTONWORKS
DATA
PLATFORM
(HDP)
OS
Cloud
VM
Appliance
© Hortonworks Inc. 2012 Page 14

15. Key 2013 “Enterprise Hadoop” Initiatives
Invest In:
Hive / “Stinger”
Interactive Query
– Platform Services
Ambari HBase – DR, Snapshot, …
Manage & Operate Online Data
OPERATIONAL
DATA
SERVICES
SERVICES
HADOOP
CORE
– Data Services
PLATFORM
SERVICES
– In support of Refine,
“Knox” HORTONWORKS
“Herd” Explore, Enrich
Secure Access
DATA
PLATFORM
(HDP)
Data Integration
– Operational Services
“Continuum” – Manageability,
Biz Continuity
Security, …
© Hortonworks Inc. 2012 Page 15

16. Hive/Stinger: Interactive Query
Near-realtime queries in good old Hive…
© Hortonworks Inc. 2012 Page 16

17. Top BI Vendors Support Hive Today
© Hortonworks Inc. 2012 Page 17

18. Goal: Enhance Hive for BI Use Cases
Enterprise Reports Parameterized Reports
Dashboard / Scorecard
Data Mining Visualization
More SQL
&
Better Performance
Batch Interactive
© Hortonworks Inc. 2012 Page 18

19. Differing Needs For Scale / Interaction
Interactivity Scalability and Reliability
is key are key
Non-
Interactive Batch
Interactive
• Parameterized • Data preparation • Operational batch
Reports • Incremental batch processing
• Drilldown processing • Enterprise
• Visualization • Dashboards / Reports
• Exploration Scorecards • Data Mining
5s – 1m 1m – 1h 1h+
Data Size
© Hortonworks Inc. 2012 Page 19

20. Stinger: Make Hive Best for All Needs
Non-
Interactive Batch
Interactive
• Parameterized • Data preparation • Operational batch
Reports • Incremental batch processing
• Drilldown processing • Enterprise
• Visualization • Dashboards / Reports
• Exploration Scorecards • Data Mining
5s – 1m 1m – 1h 1h+
Data Size
Improve Latency & Throughput Extend Deep Analytical Ability
• Query engine improvements • Analytics functions
• New “Optimized RCFile” column store • Improved SQL coverage
• Next-gen runtime (elim’s M/R latency) • Continued focus on core Hive use cases
© Hortonworks Inc. 2012 Page 20

21. Analytic Function Use Cases
• OVER
– Rankings, top 10, bottom 10
– Running balances
– Statistics within time windows (e.g. last 3 months, last 6 months)
• LEAD / LAG
– Trend identification
– Sessionization
– Forecasting / prediction
• Distributions
– Histograms and bucketing
• Good for Enterprise Reports, Dashboards, Data
Mining and Business Processing.
© Hortonworks Inc. 2012 Page 21

22. Stinger 2013 Roadmap Summary
• HDP 1.x (aka Hadoop 1.x …)
– Additional SQL Types
– SQL Analytic Functions (OVER, Subqueries in WHERE, etc.)
– Modern Optimized Column Store (ORC file)
– Hive Query Enhancements
– Startup time, star joins, optimize M/R DAGs, vectorization, etc.
• HDP 2.x (aka Hadoop 2.x …)
– Features in HDP 1.3 & 1.4
– Next-gen runtime that eliminates startup time
– Persistent function registry
– Other features
© Hortonworks Inc. 2012 Page 22

23. Questions?
© Hortonworks Inc. 2012 Page 23