2 Using Oracle Big Data SQL for Data Access
Oracle Big Data SQL enables you to query diverse data sources using the full power of Oracle SQL SELECT statements.
This chapter describes how to create Oracle Big Data SQL enabled external tables over data from Hive, Hadoop, Apache Kafka, Oracle NoSQL Database, and object stores.
2.1 Creating External Tables
Oracle Big Data SQL enables you to query external data through Oracle Big Data SQL enabled external tables from the Oracle Database using the full power of Oracle SQL SELECT statements. It also enables you to write queries that join Oracle tables and that external data, leverage robust Oracle Database security features, and take advantage of advanced SQL capabilities like analytic functions, JSON handling, and others.
The section contains the following topics:
2.1.1 About the SQL CREATE TABLE Statement
The SQL CREATE TABLE statement has a clause specifically for creating external tables. The information that you provide in this clause enables the access driver to read data from an external source and prepare the data for the external table.
2.1.1.1 Basic Syntax
The following is the basic syntax of the CREATE TABLE statement for external tables:
CREATE TABLE table_name (column_name datatype, column_name datatype[,...]) ORGANIZATION EXTERNAL (external_table_clause);
You specify the column names and data types the same as for any other table. ORGANIZATION EXTERNAL identifies the table as an external table.
The external_table_clause identifies the access driver and provides the information that it needs to load the data. See "About the External Table Clause".
2.1.1.2 About the External Table Clause
CREATE TABLE ORGANIZATION EXTERNAL statement takes the external table clause as its argument. The external table clause has the following subclauses:
TYPE
The TYPE clause identifies the access driver. The type of access driver determines how the other parts of the external table definition are interpreted.
Specify one of the following values for Oracle Big Data SQL:
-
ORACLE_HDFS: Accesses files in an HDFS directory. -
ORACLE_HIVE: Accesses a Hive table. -
ORACLE_BIGDATA: Accesses files in an object store.
Note:
The ORACLE_DATAPUMP and ORACLE_LOADER access drivers are not associated with Oracle Big Data SQL.
DEFAULT DIRECTORY
The DEFAULT DIRECTORY clause identifies an Oracle Database directory object. The directory object identifies an operating system directory with files that the external table reads and writes.
ORACLE_HDFS, ORACLE_BIGDATA, and ORACLE_HIVE use the default directory solely to write log files on the Oracle Database system.
LOCATION
The LOCATION clause identifies the data source.
REJECT LIMIT
The REJECT LIMIT clause limits the number of conversion errors permitted during a query of the external table before Oracle Database stops the query and returns an error.
Any processing error that causes a row to be rejected counts against the limit. The reject limit applies individually to each parallel query (PQ) process. It is not the total of all rejected rows for all PQ processes.
ACCESS PARAMETERS
The ACCESS PARAMETERS clause provides information that the access driver needs to load the data correctly into the external table. .
2.1.2 Creating an Oracle External Table for Hive Data
You can leverage Hive metadata when creating your Oracle Big Data SQL external tables.
External table creation is a technique to access data not only data in HDFS, but also data in other storage types, including Oracle NoSQL Database, Apache Kafka, HBase, and object stores.
To enable Oracle Big Data SQL to query Hive data, you must first define an Oracle external table for your Hive data. There are a number of tools available to help you create the Oracle external table definition. These tools leverage the underlying hive metadata, making it easier to create the tables.
The external table provides a level of abstraction. The underlying partitions or file type may change, but the Oracle external table definition can remain the same. It automatically picks up these changes when you query the table.
As part of an external table definition, you specify the table columns and their data types as well as a pointer to the source table in Hive. The rest of the metadata is derived at query execution time, including the data location, the file type and partitioning information.
-
DBMS_HADOOP
DBMS_HADOOP is a PL/SQL package that contains the
CREATE_EXTDDL_FOR_HIVEprocedure. This procedure generates the DDL to create an Oracle external table for a given Hive table. You can optionally edit the text of the generated DDL before execution in order to customize the external table properties.This procedure enables you to easily automate the definition of many tables at one time. -
The Big Data SQL wizard in Oracle SQL Developer
The most recent versions of the free Oracle SQL Developer tool include a Big Data SQL wizard that guides you easily through the process of creating Oracle external table definitions.
If you have a configured Hive connection in Oracle SQL Developer, then in the Connections navigator, drill down from the connection entry point to a Hive table and do the following:
-
Right-click on the table icon and select Use in Oracle Big Data SQL...
-
When prompted, select an Oracle Database connection for the import of the Hive table.
-
Select an Oracle Big Data SQL-enabled target database.
-
In the Create Table dialog, check over the current configuration for columns, external table properties, and storage. Modify as needed. You can also preview the text of the DDL that will be generated.
-
Click OK when you are satisfied with the table definition. The wizard will create the external table at the designated location.
-
-
The Oracle SQL Developer Data Modeler
This is free graphical design tool that you can use to connect to a Hive metastore and generate an external table. You can select and import one or multiple Hive tables, modify table properties as needed, and then generate the DDL that you can copy into an SQL Worksheet and then run in order to create an Oracle external table. Although the Data Modeler is a more complex tool to use than the other options, its advantage is that you can use it to work on multiple Hive tables
See Oracle SQL Developer & Data Modeler Support for Oracle Big Data SQL in the Oracle Blog space for a demonstration of how to use the Data Modeler.
See Also:
For instructions on how to install Oracle SQL Developer and connect to Hive in order to create external tables, see Using Oracle SQL Developer to Connect to Hive2.1.2.1 Obtaining Information About a Hive Table
The DBMS_HADOOP PL/SQL package contains a function named CREATE_EXTDDL_FOR_HIVE. It returns the data dictionary language (DDL) to create an external table for accessing a Hive table. This function requires you to provide basic information about the Hive table:
-
Name of the Hadoop cluster
-
Name of the Hive database
-
Name of the Hive table
-
Whether the Hive table is partitioned
You can obtain this information by querying the ALL_HIVE_TABLES data dictionary view. It displays information about all Hive tables that you can access from Oracle Database.
This example shows that the current user has access to an unpartitioned Hive table named RATINGS_HIVE_TABLE in the default database. A user named JDOE is the owner.
SQL> SELECT cluster_id, database_name, owner, table_name, partitioned FROM all_hive_tables; CLUSTER_ID DATABASE_NAME OWNER TABLE_NAME PARTITIONED ------------ -------------- -------- ------------------ -------------- hadoop1 default jdoe ratings_hive_table UN-PARTITIONED
See Also:
2.1.2.2 Using the CREATE_EXTDDL_FOR_HIVE Function
With the information from the data dictionary, you can use the CREATE_EXTDDL_FOR_HIVE function of DBMS_HADOOP. This example specifies a database table name of RATINGS_DB_TABLE in the current schema. The function returns the text of the CREATE TABLE command in a local variable named DDLout, but does not execute it.
DECLARE
DDLout VARCHAR2(4000);
BEGIN
dbms_hadoop.create_extddl_for_hive(
CLUSTER_ID=>'hadoop1',
DB_NAME=>'default',
HIVE_TABLE_NAME=>'ratings_hive_table',
HIVE_PARTITION=>FALSE,
TABLE_NAME=>'ratings_db_table',
PERFORM_DDL=>FALSE,
TEXT_OF_DDL=>DDLout
);
dbms_output.put_line(DDLout);
END;
/When this procedure runs, the PUT_LINE function displays the CREATE TABLE command:
CREATE TABLE ratings_db_table (
c0 VARCHAR2(4000),
c1 VARCHAR2(4000),
c2 VARCHAR2(4000),
c3 VARCHAR2(4000),
c4 VARCHAR2(4000),
c5 VARCHAR2(4000),
c6 VARCHAR2(4000),
c7 VARCHAR2(4000))
ORGANIZATION EXTERNAL
(TYPE ORACLE_HIVE DEFAULT DIRECTORY DEFAULT_DIR
ACCESS PARAMETERS
(
com.oracle.bigdata.cluster=hadoop1
com.oracle.bigdata.tablename=default.ratings_hive_table
)
) PARALLEL 2 REJECT LIMIT UNLIMITEDYou can capture this information in a SQL script, and use the access parameters to change the Oracle table name, the column names, and the data types as desired before executing it. You might also use access parameters to specify a date format mask.
The ALL_HIVE_COLUMNS view shows how the default column names and data types are derived. This example shows that the Hive column names are C0 to C7, and that the Hive STRING data type maps to VARCHAR2(4000):
SQL> SELECT table_name, column_name, hive_column_type, oracle_column_type FROM all_hive_columns; TABLE_NAME COLUMN_NAME HIVE_COLUMN_TYPE ORACLE_COLUMN_TYPE --------------------- ------------ ---------------- ------------------ ratings_hive_table c0 string VARCHAR2(4000) ratings_hive_table c1 string VARCHAR2(4000) ratings_hive_table c2 string VARCHAR2(4000) ratings_hive_table c3 string VARCHAR2(4000) ratings_hive_table c4 string VARCHAR2(4000) ratings_hive_table c5 string VARCHAR2(4000) ratings_hive_table c6 string VARCHAR2(4000) ratings_hive_table c7 string VARCHAR2(4000) 8 rows selected.
See Also:
2.1.2.3 Using Oracle SQL Developer to Connect to Hive
Oracle SQL Developer provides methods to connect to a Hive metastore and create Oracle external tables over Hive.
Follow these steps to set up Oracle SQL Developer to work with Oracle Big Data SQL.
-
Install Oracle SQL Developer
-
Download the Hive JDBC Drivers
-
Add the new Hive JDBC Drivers to Oracle SQL Developer
-
Create a database connection to Hive.
Installing Oracle SQL Developer
Install Oracle SQL Developer 4.2 or greater. Starting with this version, support is included for Copy To Hadoop, a useful Oracle Big Data SQL tool for off-loading Oracle Database tables to HDFS.
The installation is simple. Just download the package and extract it.
-
Go to the Oracle SQL Developer download site on the Oracle Technology Network (OTN).
-
Accept the license agreement and download the version that is appropriate for your platform.
For most users, Windows 64–bit with JDK 8 included is the correct choice.
-
Extract the downloaded ZIP file to your local drive.
You can extract to any folder name.
See Installing and Getting Started with SQL Developer in the Oracle SQL Developer User’s Guide for further installation and configuration details.
Downloading and Installing the Hive JDBC Drivers for Cloudera Enterprise
To connect Oracle SQL Developer to Hive in the Hadoop environment, you need to download and install the Hive JDBC drivers for Cloudera Enterprise. These drivers are not included in the Oracle SQL Developer download package.
Note for HDP Users:
At this time, SQL Developer 4.2 requires the Cloudera JDBC drivers for Hive. However, these drivers appear to work against Hortonworks clusters as well. HDP users should test to determine if these drivers meet their needs.-
Download the latest Cloudera JDBC drivers for Hive from the Cloudera website to any local directory.
You can search for “
cloudera hive jdbc drivers download” on the Cloudera website to locate the available driver packages.You are prompted to select the driver version, OS, and OS version (32/64 bit). At this time, the latest drive version is 2.5.18. You can choose the newest version available.
-
Unzip the archive:
unzip hive_jdbc_<version>.zip -
View the extracted content. Notice that under the top-level folder there are multiple ZIP files. Each is for a different JDBC version. For this setup, only JBDC 4.0 is usable. Select the JDBC4_ ZIP file (
JDBC4_<version>.zip).Important:
Choose only the JDBC4_ ZIP file, which contains the drivers for JDBC 4.0. This is the only compatible version. The drivers in other packages, such asJDBC41_*, are not compatible with SQL Developer 4.2 and will return errors upon connection attempts. -
Unzip the JDBC4 archive to a target directory that is accessible to Oracle SQL Developer, for example,
./home/oracle/jdbc:# unzip Cloudera_HiveJDBC4_<version>.zip -d /home/oracle/jdbc/The extracted content should be similar to this:
Cloudera_HiveJDBC4_2.5.18.1050\Cloudera-JDBC-Driver-for-Apache-Hive-Install-Guide.pdf Cloudera_HiveJDBC4_2.5.18.1050\Cloudera-JDBC-Driver-for-Apache-Hive-Release-Notes.pdf Cloudera_HiveJDBC4_2.5.18.1050\commons-codec-1.3.jar Cloudera_HiveJDBC4_2.5.18.1050\commons-logging-1.1.1.jar Cloudera_HiveJDBC4_2.5.18.1050\HiveJDBC4.jar Cloudera_HiveJDBC4_2.5.18.1050\hive_metastore.jar Cloudera_HiveJDBC4_2.5.18.1050\hive_service.jar Cloudera_HiveJDBC4_2.5.18.1050\httpclient-4.1.3.jar Cloudera_HiveJDBC4_2.5.18.1050\httpcore-4.1.3.jar Cloudera_HiveJDBC4_2.5.18.1050\libfb303-0.9.0.jar Cloudera_HiveJDBC4_2.5.18.1050\libthrift-0.9.0.jar Cloudera_HiveJDBC4_2.5.18.1050\log4j-1.2.14.jar Cloudera_HiveJDBC4_2.5.18.1050\out.txt Cloudera_HiveJDBC4_2.5.18.1050\ql.jar Cloudera_HiveJDBC4_2.5.18.1050\slf4j-api-1.5.11.jar Cloudera_HiveJDBC4_2.5.18.1050\slf4j-log4j12-1.5.11.jar Cloudera_HiveJDBC4_2.5.18.1050\TCLIServiceClient.jar Cloudera_HiveJDBC4_2.5.18.1050\zookeeper-3.4.6.jar
Add the new Hive JDBC Drivers to Oracle SQL Developer
Next, start up SQL Developer and copy all of the extracted driver files into “Third Party JDBC Drivers” in the Preferences window.
- Start SQL Developer.
-
In the SQL Developer menu bar, select Tools>Preferences.
-
In the file explorer of the Preferences window, expand Database and then click Third Party JDBC Drivers.
-
Click Add Entry.
-
Navigate to the folder where you sent the files extracted from
Cloudera_HiveJDBC4_<version>.zip. Copy all of the JAR files from the ZIP extraction into this window and then click OK. -
Restart Oracle SQL Developer.
Create a Database Connection to Hive
After the drivers are installed, you can create a connection to Hiveserver2.
If you are creating a Kerberos-secured connection, you will need a user ID, the Kerberos connection parameters, and the number of the port where Hiveserver2 is running on the Hadoop system (typically, port 10000). A keytab must exist for the user.
If you not using Kerberos, you will need a user ID (the oracle user or a user with equivalent privileges), the account password, and the Hiveserver2 port number.
See Create/Edit/Select Database Connection in the Oracle SQL Developer User’s Guide for a explanation of the fields in the Oracle and Hive tabs in the New/Select Database Connection dialog.
2.1.2.4 Developing a CREATE TABLE Statement for ORACLE_HIVE
Whichever method you use to create an Oracle external table over Hive (DBMS_HADOOP, Oracle SQL Developer Data Modeler, Oracle Big Data Wizard in Oracle SQL Developer, or manual coding), you may need to set some access parameters to modify the default behavior of ORACLE_HIVE.
Note:
Do not include the LOCATION clause with ORACLE_HIVE. It raises an error. The data is stored in Hive, and the access parameters and the metadata store provide the necessary information.
2.1.2.4.1 Using the Default ORACLE_HIVE Settings
The following statement creates an external table named ORDER to access Hive data:
CREATE TABLE order (cust_num VARCHAR2(10),
order_num VARCHAR2(20),
description VARCHAR2(100),
order_total NUMBER (8,2))
ORGANIZATION EXTERNAL (TYPE oracle_hive);Because no access parameters are set in the statement, the ORACLE_HIVE access driver uses the default settings to do the following:
-
Connects to the default Hadoop cluster.
-
Uses a Hive table named
order. An error results if the Hive table does not have fields namedCUST_NUM,ORDER_NUM,DESCRIPTION, andORDER_TOTAL. -
Sets the value of a field to
NULLif there is a conversion error, such as aCUST_NUMvalue longer than 10 bytes.
2.1.2.4.2 Overriding the Default ORACLE_HIVE Settings
You can set properties in the ACCESS PARAMETERS clause of the external table clause, which override the default behavior of the access driver. The following clause includes the com.oracle.bigdata.overflow access parameter. When this clause is used in the previous example, it truncates the data for the DESCRIPTION column that is longer than 100 characters, instead of throwing an error:
(TYPE oracle_hive
ACCESS PARAMETERS (
com.oracle.bigdata.overflow={"action:"truncate", "col":"DESCRIPTION""} ))The next example sets most of the available parameters for ORACLE_HIVE:
CREATE TABLE order (cust_num VARCHAR2(10),
order_num VARCHAR2(20),
order_date DATE,
item_cnt NUMBER,
description VARCHAR2(100),
order_total (NUMBER(8,2)) ORGANIZATION EXTERNAL
(TYPE oracle_hive
ACCESS PARAMETERS (
com.oracle.bigdata.tablename: order_db.order_summary
com.oracle.bigdata.colmap: {"col":"ITEM_CNT", \
"field":"order_line_item_count"}
com.oracle.bigdata.overflow: {"action":"TRUNCATE", \
"col":"DESCRIPTION"}
com.oracle.bigdata.erroropt: [{"action":"replace", \
"value":"INVALID_NUM" , \
"col":["CUST_NUM","ORDER_NUM"]} ,\
{"action":"reject", \
"col":"ORDER_TOTAL}
))The parameters make the following changes in the way that the ORACLE_HIVE access driver locates the data and handles error conditions:
-
com.oracle.bigdata.tablename: Handles differences in table names.ORACLE_HIVElooks for a Hive table namedORDER_SUMMARYin theORDER.DBdatabase. -
com.oracle.bigdata.colmap: Handles differences in column names. The HiveORDER_LINE_ITEM_COUNTfield maps to the OracleITEM_CNTcolumn. -
com.oracle.bigdata.overflow: Truncates string data. Values longer than 100 characters for theDESCRIPTIONcolumn are truncated. -
com.oracle.bigdata.erroropt: Replaces bad data. Errors in the data forCUST_NUMorORDER_NUMset the value toINVALID_NUM.
See Also:
The section CREATE TABLE ACCESS PARAMETERS Clause provides the complete list of access parameters forORACLE_HIVE, ORACLE_HDFS, and ORACLE_BIGDATA.
2.1.2.5 Hive to Oracle Data Type Conversions
When the access driver loads data into an external table, it verifies that the Hive data can be converted to the data type of the target column. If they are incompatible, then the access driver returns an error. Otherwise, it makes the appropriate data conversion.
Hive typically provides a table abstraction layer over data stored elsewhere, such as in HDFS files. Hive uses a serializer/deserializer (SerDe) to convert the data as needed from its stored format into a Hive data type. The access driver then converts the data from its Hive data type to an Oracle data type. For example, if a Hive table over a text file has a BIGINT column, then the SerDe converts the data from text to BIGINT. The access driver then converts the data from BIGINT (a Hive data type) to NUMBER (an Oracle data type).
Performance is better when one data type conversion is performed instead of two. The data types for the fields in the HDFS files should therefore indicate the data that is actually stored on disk. For example, JSON is a clear text format, therefore all data in a JSON file is text. If the Hive type for a field is DATE, then the SerDe converts the data from string (in the data file) to a Hive date. Then the access driver converts the data from a Hive date to an Oracle date. However, if the Hive type for the field is string, then the SerDe does not perform a conversion, and the access driver converts the data from string to an oracle date. Queries against the external table are faster in the second example, because the access driver performs the only data conversion.
The table below identifies the data type conversions that ORACLE_HIVE can make when loading data into an external table.
Table 2-1 Supported Hive to Oracle Data Type Conversions
| Hive Data Type | VARCHAR2, CHAR, NCHAR2, NCHAR, CLOB | NUMBER, FLOAT, BINARY_NUMBER, BINARY_FLOAT | BLOB | RAW | DATE, TIMESTAMP, TIMESTAMP WITH TZ, TIMESTAMP WITH LOCAL TZ | INTERVAL YEAR TO MONTH, INTERVAL DAY TO SECOND |
|---|---|---|---|---|---|---|
|
INT SMALLINT TINYINT BIGINT |
yes |
yes |
yes |
yes |
no |
no |
|
DOUBLE FLOAT |
yes |
yes |
yes |
yes |
no |
no |
|
DECIMAL |
yes |
yes |
no |
no |
no |
no |
|
BOOLEAN |
yesFoot 1 |
yesFoot 2 |
yesFoot 2 |
yes |
no |
no |
|
BINARY |
yes |
no |
yes |
yes |
no |
no |
|
STRING |
yes |
yes |
no |
no |
yes |
yes |
|
TIMESTAMP |
yes |
no |
no |
no |
yes |
no |
|
STRUCT ARRAY UNIONTYPE MAP |
yes |
no |
no |
no |
no |
no |
Footnote 1
FALSE maps to the string FALSE, and TRUE maps to the string TRUE.
Footnote 2
FALSE maps to 0, and TRUE maps to 1.
2.1.3 Creating an Oracle External Table for Oracle NoSQL Database
You can use the ORACLE_HIVE access driver to access data stored in Oracle NoSQL Database. However, you must first create a Hive external table that accesses the KVStore. Then you can create an external table in Oracle Database over it, similar to the process described in "Creating an Oracle External Table for Hive Data".
This section contains the following topics:
2.1.3.1 Creating a Hive External Table for Oracle NoSQL Database
To provide access to the data in Oracle NoSQL Database, you create a Hive external table over the Oracle NoSQL table. Oracle Big Data SQL provides a StorageHandler named oracle.kv.hadoop.hive.table.TableStorageHandler that enables Hive to read the Oracle NoSQL Database table format.
The following is the basic syntax of a Hive CREATE TABLE statement for a Hive external table over an Oracle NoSQL table:
CREATE EXTERNAL TABLE tablename colname coltype[, colname coltype,...] STORED BY 'oracle.kv.hadoop.hive.table.TableStorageHandler' TBLPROPERTIES ( "oracle.kv.kvstore" = "database", "oracle.kv.hosts" = "nosql_node1:port[, nosql_node2:port...]", "oracle.kv.hadoop.hosts" = "hadoop_node1[,hadoop_node2...]", "oracle.kv.tableName" = "table_name");
Hive CREATE TABLE Parameters
- tablename
-
The name of the Hive external table being created.
This table name will be used in SQL queries issued in Oracle Database, so choose a name that is appropriate for users. The name of the external table that you create in Oracle Database must be identical to the name of this Hive table.
Table, column, and field names are case insensitive in Oracle NoSQL Database, Apache Hive, and Oracle Database.
- colname coltype
-
The names and data types of the columns in the Hive external table. See Table 2-2 for the data type mappings between Oracle NoSQL Database and Hive.
Hive CREATE TABLE TBLPROPERTIES Clause
- oracle.kv.kvstore
-
The name of the KVStore. Only upper- and lowercase letters and digits are valid in the name.
- oracle.kv.hosts
-
A comma-delimited list of host names and port numbers in the Oracle NoSQL Database cluster. Each string has the format hostname:port. Enter multiple names to provide redundancy in the event that a host fails.
- oracle.kv.hadoop.hosts
-
A comma-delimited list of all host names in the Hadoop cluster with Oracle Big Data SQL enabled.
- oracle.kv.tableName
-
The name of the table in Oracle NoSQL Database that stores the data for this Hive external table.
See Also:
Apache Hive Language Manual DDL at
2.1.3.2 Creating the Oracle Database Table for Oracle NoSQL Data
Use the following syntax to create an external table in Oracle Database that can access the Oracle NoSQL data through a Hive external table:
CREATE TABLE tablename(colname colType[,colname colType...]) ORGANIZATION EXTERNAL (TYPE ORACLE_HIVE DEFAULT DIRECTORY directory ACCESS PARAMETERS (access parameters) ) REJECT LIMIT UNLIMITED;
In this syntax, you identify the column names and data types. For more about this syntax, see "About the SQL CREATE TABLE Statement".
2.1.3.3 About Oracle NoSQL to Oracle Database Type Mappings
When Oracle Big Data SQL retrieves data from Oracle NoSQL Database, the data is converted twice to another data type:
-
To a Hive data type when the data is read into the columns of the Hive external table.
-
To an Oracle data type when the data is read into the columns of an Oracle Database external table.
In order to execute a Big Data SQL query against data stored in an Oracle NoSQL Database table, a Hive external table must first be created with a schema mapped from the schema of the desired Oracle NoSQL Database table.Table 2-2 identifies the supported data types Oracle NoSQL Database table API and their mappings to Hive.
Table 2-2 Mapping Hive Data Types to the NoSQL Database Table API Data Model
| Oracle NoSQL Database Table API | Hive |
|---|---|
|
FieldDef.Type.STRING |
STRING |
|
FieldDef.Type.BOOLEAN |
BOOLEAN |
|
FieldDef.Type.BINARY |
BINARY |
|
FieldDef.Type.FIXED_BINARY |
BINARY |
|
FieldDef.Type.INTEGER |
INT |
|
FieldDef.Type.LONG |
BIGINT |
|
FieldDef.Type.FLOAT |
FLOAT |
|
FieldDef.Type.DOUBLE |
DOUBLE |
|
FieldDef.Type.ENUM |
STRING |
|
FieldDef.Type.ARRAY |
ARRAY |
|
FieldDef.Type.MAP |
MAP<STRING, data_type> |
|
FieldDef.Type.RECORD |
STRUCT<col_name : data_type, ...> |
Note:
To complete this mapping a corresponding Oracle Database external table must be created with a schema mapped from the schema of the Hive table.
Also note that the following Hive data types are not applicable to the mapping of Oracle NoSQL data types to Oracle Database data types: VARCHAR, CHAR, TINYINT, SMALLINT, DECIMAL, TIMESTAMP, DATE, UNION TYPE.
See Also:
Hive to Oracle Data Type Conversions provides details on Hive to Oracle Database data type mappings.Predicate Pushdown in Oracle Big Data SQL requires that certain mappings between Hive Datatypes and Oracle Datatypes be present. See About Predicate Push Down.
2.1.3.4 Example of Accessing Data in Oracle NoSQL Database
This example uses the sample data provided with the Oracle NoSQL Database software:
2.1.3.4.1 Creating the Oracle NoSQL Database Example Table
Verify that the following files reside in the examples/hadoop/table directory:
create_vehicle_table.kvs CountTableRows.java LoadVehicleTable.java
This example runs on a Hadoop cluster node named some1node07 and uses a KVStore named SOME1KV.
To create and populate the sample table in Oracle NoSQL Database:
-
Open a connection to an Oracle NoSQL Database node on your Hadoop cluster.
-
Create a table named
vehicleTable. The following example uses theloadcommand to run the commands increate_vehicle_table.kvs:$ cd NOSQL_HOME $ java -jar lib/kvcli.jar -host some1node07 -port 5000 \ load -file examples/hadoop/table/create_vehicle_table.kvs -
Compile
LoadVehicleTable.java:$ javac -cp examples:lib/kvclient.jar examples/hadoop/table/LoadVehicleTable.java
-
Execute the
LoadVehicleTableclass to populate the table:$ java -cp examples:lib/kvclient.jar hadoop.table.LoadVehicleTable -host some1node07 -port 5000 -store SOME1KV {"type":"auto","make":"Chrysler","model":"PTCruiser","class":"4WheelDrive","colo r":"white","price":20743.240234375,"count":30} {"type":"suv","make":"Ford","model":"Escape","class":"FrontWheelDrive","color":" . . . 10 new records added
The vehicleTable table contains the following fields:
Table 2-3 Fields in the vehicleTable Example
| Field Name | Data Type |
|---|---|
|
type |
STRING |
|
make |
STRING |
|
model |
STRING |
|
class |
STRING |
|
color |
STRING |
|
price |
DOUBLE |
|
count |
INTEGER |
2.1.3.4.2 Creating the Example Hive Table for vehicleTable
The following example creates a Hive table named VEHICLES that accesses vehicleTable in the SOME1KV KVStore. In this example, the system is configured with a Hadoop cluster in the first six servers (some1node01 to some1node06) and an Oracle NoSQL Database cluster in the next three servers (some1node07 to some1node09).
CREATE EXTERNAL TABLE IF NOT EXISTS vehicles
(type STRING,
make STRING,
model STRING,
class STRING,
color STRING,
price DOUBLE,
count INT)
COMMENT 'Accesses data in vehicleTable in the SOME1KV KVStore'
STORED BY 'oracle.kv.hadoop.hive.table.TableStorageHandler'
TBLPROPERTIES
("oracle.kv.kvstore" = "SOME1KV",
"oracle.kv.hosts" = "some1node07.example.com:5000,some1node08.example.com:5000",
"oracle.kv.hadoop.hosts" = "some1node01.example.com,some1node02.example.com,some1node03.example.com,some1node04.example.com,some1node05.example.com,some1node06.example.com",
"oracle.kv.tableName" = "vehicleTable");The DESCRIBE command lists the columns in the VEHICLES table:
hive> DESCRIBE vehicles;
OK
type string from deserializer
make string from deserializer
model string from deserializer
class string from deserializer
color string from deserializer
price double from deserializer
count int from deserializerA query against the Hive VEHICLES table returns data from the Oracle NoSQL vehicleTable table:
hive> SELECT make, model, class FROM vehicletable WHERE type='truck' AND color='red' ORDER BY make, model; Total MapReduce jobs = 1 Launching Job 1 out of 1 Number of reduce tasks determined at compile time: 1 . . . Chrysler Ram1500 RearWheelDrive Chrysler Ram2500 FrontWheelDrive Ford F150 FrontWheelDrive Ford F250 RearWheelDrive Ford F250 AllWheelDrive Ford F350 RearWheelDrive GM Sierra AllWheelDrive GM Silverado1500 RearWheelDrive GM Silverado1500 AllWheelDrive
2.1.3.4.3 Creating the Oracle Table for VEHICLES
After you create the Hive table, the metadata is available in the Oracle Database static data dictionary views. The following SQL SELECT statement returns information about the Hive table created in the previous topic:
SQL> SELECT table_name, column_name, hive_column_type FROM all_hive_columns WHERE table_name='vehicles'; TABLE_NAME COLUMN_NAME HIVE_COLUMN_TYPE --------------- ------------ ---------------- vehicles type string vehicles make string vehicles model string vehicles class string vehicles color string vehicles price double vehicles count int
The next SQL CREATE TABLE statement generates an external table named VEHICLES over the Hive VEHICLES table, using the ORACLE_HIVE access driver. The name of the table in Oracle Database must be identical to the name of the table in Hive. However, both Oracle NoSQL Database and Oracle Database are case insensitive.
CREATE TABLE vehicles
(type VARCHAR2(10), make VARCHAR2(12), model VARCHAR2(20),
class VARCHAR2(40), color VARCHAR2(20), price NUMBER(8,2),
count NUMBER)
ORGANIZATION EXTERNAL
(TYPE ORACLE_HIVE DEFAULT DIRECTORY DEFAULT_DIR
ACCESS PARAMETERS
(com.oracle.bigdata.debug=true com.oracle.bigdata.log.opt=normal))
REJECT LIMIT UNLIMITED;This SQL SELECT statement retrieves all rows for red trucks from vehicleTable in Oracle NoSQL Database:
SQL> SELECT make, model, class FROM vehicles WHERE type='truck' AND color='red' ORDER BY make, model; MAKE MODEL CLASS ------------ -------------------- --------------------- Chrysler Ram1500 RearWheelDrive Chrysler Ram2500 FrontWheelDrive Ford F150 FrontWheelDrive Ford F250 AllWheelDrive Ford F250 RearWheelDrive Ford F350 RearWheelDrive GM Sierra AllWheelDrive GM Silverado1500 RearWheelDrive GM Silverado1500 4WheelDrive GM Silverado1500 AllWheelDrive
2.1.4 Creating an Oracle External Table for Apache HBase
You can also use the ORACLE_HIVE access driver to access data stored in Apache HBase. However, you must first create a Hive external table that accesses the HBase table. Then you can create an external table in Oracle Database over it. The basic steps are the same as those described in "Creating an Oracle External Table for Oracle NoSQL Database".
2.1.4.1 Creating a Hive External Table for HBase
To provide access to the data in an HBase table, you create a Hive external table over it. Apache provides a storage handler and a SerDe that enable Hive to read the HBase table format.
The following is the basic syntax of a Hive CREATE TABLE statement for an external table over an HBase table:
CREATE EXTERNAL TABLE tablename colname coltype[, colname coltype,...] ROW FORMAT SERDE 'org.apache.hadoop.hive.hbase.HBaseSerDe' STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler' WITH SERDEPROPERTIES ( 'serialization.format'='1', 'hbase.columns.mapping'=':key,value:key,value:
See Also:
-
Apache Hive Language Manual DDL at
-
Hive HBase Integration at
https://cwiki.apache.org/confluence/display/Hive/HBaseIntegration#HBaseIntegration-StorageHandlers -
Class
HBaseSerDein the Apache Hive Javadocs athttps://hive.apache.org/javadoc.html
2.1.4.2 Creating the Oracle Database Table for HBase
Use the following syntax to create an external table in Oracle Database that can access the HBase data through a Hive external table:
CREATE TABLE tablename(colname colType[,colname colType...]) ORGANIZATION EXTERNAL (TYPE ORACLE_HIVE DEFAULT DIRECTORY DEFAULT_DIR ACCESS PARAMETERS (access parameters) ) REJECT LIMIT UNLIMITED;
In this syntax, you identify the column names and data types. To specify the access parameters, see "About the SQL CREATE TABLE Statement".
2.1.5 Creating an Oracle External Table for HDFS Files
The ORACLE_HDFS access driver enables you to access many types of data that are stored in HDFS, but which do not have Hive metadata. You can define the record format of text data, or you can specify a SerDe for a particular data format.
You must create the external table for HDFS files manually, and provide all the information the access driver needs to locate the data, and parse the records and fields. The following are some examples of CREATE TABLE ORGANIZATION EXTERNAL statements.
2.1.5.1 Using the Default Access Parameters with ORACLE_HDFS
The following statement creates a table named ORDER to access the data in all files stored in the /usr/cust/summary directory in HDFS:
CREATE TABLE ORDER (cust_num VARCHAR2(10),
order_num VARCHAR2(20),
order_total NUMBER (8,2))
ORGANIZATION EXTERNAL
( TYPE oracle_hdfs
DEFAULT DIRECTORY DEFAULT_DIR
)
LOCATION ('hdfs:/usr/cust/summary/*');Because no access parameters are set in the statement, the ORACLE_HDFS access driver uses the default settings to do the following:
-
Connects to the default Hadoop cluster.
-
Reads the files as delimited text, and the fields as type
STRING. -
Assumes that the number of fields in the HDFS files match the number of columns (three in this example).
-
Assumes the fields are in the same order as the columns, so that
CUST_NUMdata is in the first field,ORDER_NUMdata is in the second field, andORDER_TOTALdata is in the third field. -
Rejects any records in which the value causes a data conversion error: If the value for
CUST_NUMexceeds 10 characters, the value forORDER_NUMexceeds 20 characters, or the value ofORDER_TOTALcannot be converted toNUMBER.
See Also:
The section CREATE TABLE ACCESS PARAMETERS Clause provides the complete list of access parameters forORACLE_HIVE, ORACLE_HDFS, and ORACLE_BIGDATA.
2.1.5.2 ORACLE_HDFS LOCATION Clause
The LOCATION clause for ORACLE_HDFS contains a comma-separated list of file locations. The files must reside in the HDFS file system on the default cluster.
A location can be any of the following:
-
A fully qualified HDFS directory name, such as
/user/hive/warehouse/hive_seed/hive_types.ORACLE_HDFSuses all files in the directory. -
A fully qualified HDFS file name, such as
/user/hive/warehouse/hive_seed/hive_types/hive_types.csv -
A URL for an HDFS file or a set of files, such as
hdfs:/user/hive/warehouse/hive_seed/hive_types/*. It is invalid to use the directory name alone.
The file names can contain any pattern-matching character described in Table 2-4.
Table 2-4 Pattern-Matching Characters
| Character | Description |
|---|---|
|
? |
Matches any one character |
|
* |
Matches zero or more characters |
|
[abc] |
Matches one character in the set {a, b, c} |
|
[a-b] |
Matches one character in the range {a...b}. The character must be less than or equal to b. |
|
[^a] |
Matches one character that is not in the character set or range {a}. The carat (^) must immediately follow the left bracket, with no spaces. |
|
\c |
Removes any special meaning of c. The backslash is the escape character. |
|
{ab\,cd} |
Matches a string from the set {ab, cd}. The escape character (\) removes the meaning of the comma as a path separator. |
|
{ab\,c{de\,fh} |
Matches a string from the set {ab, cde, cfh}. The escape character (\) removes the meaning of the comma as a path separator. |
2.1.5.3 Overriding the Default ORACLE_HDFS Settings
You can use many of the same access parameters with ORACLE_HDFS as ORACLE_HIVE.
2.1.5.3.1 Accessing a Delimited Text File
The following example is equivalent to the one shown in "Overriding the Default ORACLE_HIVE Settings". The external table accesses a delimited text file stored in HDFS.
CREATE TABLE taxis
(
dispatching_base_num varchar2(100),
pickup_date varchar2(100),
location_id varchar2(100)
)
ORGANIZATION EXTERNAL
(TYPE ORACLE_HDFS
DEFAULT DIRECTORY DEFAULT_DIR
ACCESS PARAMETERS
(
com.oracle.bigdata.fileformat=TEXTFILE
com.oracle.bigdata.rowformat=DELIMITED FIELDS TERMINATED BY ','
)
LOCATION ('/data/taxi-trips/')
)
REJECT LIMIT UNLIMITED;
Note that there is no colmap field, since this source contains no metadata that describes columns. The only access parameters needed in this case are fileformat and rowformat.
Tip:
Instead of a colon (:) as the delimiter between fields and values, it is better to use the equal sign (=). This is because certain tools prompt you for a parameter value when they encounter the colon delimiter.
2.1.5.3.2 Accessing Avro Container Files
The next example uses a SerDe to access Avro container files.
CREATE TABLE order (cust_num VARCHAR2(10),
order_num VARCHAR2(20),
order_date DATE,
item_cnt NUMBER,
description VARCHAR2(100),
order_total NUMBER(8,2))
ORGANIZATION EXTERNAL
(
TYPE oracle_hdfs
DEFAULT DIRECTORY DEFAULT_DIR
ACCESS PARAMETERS (
com.oracle.bigdata.rowformat: \
SERDE 'org.apache.hadoop.hive.serde2.avro.AvroSerDe'
com.oracle.bigdata.fileformat: \
INPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerInputFormat'\
OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerOutputFormat'
com.oracle.bigdata.colmap: { "col":"item_cnt", \
"field":"order_line_item_count"}
com.oracle.bigdata.overflow: {"action":"TRUNCATE", \
"col":"DESCRIPTION"}
)
LOCATION ('hdfs:/usr/cust/summary/*'));The access parameters provide the following information to the ORACLE_HDFS access driver:
-
com.oracle.bigdata.rowformat: Identifies the SerDe that the access driver needs to use to parse the records and fields. -
com.oracle.bigdata.fileformat: Identifies the Java classes that can extract records and output them in the desired format. -
com.oracle.bigdata.colmap: Handles differences in column names.ORACLE_HDFSmatchesORDER_LINE_ITEM_COUNTin the HDFS files with theITEM_CNTcolumn in the external table. -
com.oracle.bigdata.overflow: Truncates string data. Values longer than 100 characters for theDESCRIPTIONcolumn are truncated.
2.1.5.3.3 Accessing JSON Data
Oracle Big Data SQL user functionality built into Oracle SQL in order to parse data in JSON format.
Oracle SQL can parse JSON data accessible in columns (which may be external data or data stored inside the database).
For example, here is a JSON file called station_information.json, stored in HDFS.
{
"station_id":"72","name":"W 52 St & 11 Ave","short_name":"6926.01",
"lat":40.76727216,"lon":-73.99392888,"region_id":71,"rental_methods":["CREDITCARD","KEY"],
"capacity":39,"rental_url":"http://app.citibikenyc.com/S6Lr/IBV092JufD?station_id=72",
"eightd_has_key_dispenser":false}{"station_id":"79","name":"Franklin St & W Broadway",
"short_name":"5430.08","lat":40.71911552,"lon":-74.00666661,"region_id":71,
"rental_methods":["CREDITCARD","KEY"],
"capacity":33,"rental_url":"http://app.citibikenyc.com/S6Lr/IBV092JufD?station_id=79",
"eightd_has_key_dispenser":false}{"station_id":"82",
"name":"St James Pl & Pearl St","short_name":"5167.06","lat":40.71117416,
"lon":-74.00016545,"region_id":71,"rental_methods":["CREDITCARD","KEY"],
"capacity":27,"rental_url":"http://app.citibikenyc.com/S6Lr/IBV092JufD?station_id=82",
"eightd_has_key_dispenser":false
}
To query this data, do the following.
- First create an HDFS external table over this data. Add a single text as column as VARCHAR2. (For large JSON objects, you can use CLOB instead.)
CREATE TABLE bikes.stations_ext ( doc varchar2(4000) ) ORGANIZATION EXTERNAL ( TYPE ORACLE_HDFS DEFAULT DIRECTORY DEFAULT_DIR LOCATION ('/data/bike-stations') ) REJECT LIMIT UNLIMITED; - Then query the external table. Note the use of Oracle SQL's
JSON_VALUEfunction.select s.doc.station_id, s.doc.name, s.doc.lon as longitude, s.doc.lat as latitude, s.doc.capacity, s.doc.eightd_has_key_dispenser, s.doc.rental_methods, json_value(doc, '$.rental_methods[0]'), json_value(doc, '$.rental_methods[1]') from stations_exts
2.1.6 Creating an Oracle External Table for Kafka Topics
The Hive storage handler for Kafka enables Hive and Oracle Big Data SQL and Hive to query Kafka topics.
The ORACLE_HIVE access driver can access Kafka data topics. You first create a Hive external table that accesses the Kafka topics and then create an Oracle Big Data SQL table over it.
2.1.6.1 Using Oracle's Hive Storage Handler for Kafka to Create a Hive External Table for Kafka Topics
The Hive storage handler for Kafka enables Hive and Oracle Big Data SQL to query Kafka topics.
To provide access to Kafka data, you create a Hive external table over the Kafka topics. The Oracle Big Data SQL storage handler that enables Hive to read the Kafka data format is oracle.hadoop.kafka.hive.KafkaStorageHandler .
You can use this storage handler to create external Hive tables backed by data residing in Kafka. Big Data SQL can then query the Kafka data through the external Hive tables.
The Hive DDL is demonstrated by the following example, where topic1 and topic2 are two topics in Kafka broker whose keys are serialized by Kafka's String serializer and whose values are serialized by Kafka's Long serializer.
CREATE EXTERNAL TABLE test_table
row format serde ‘oracle.hadoop.kafka.hive.KafkaSerDe’
stored by 'oracle.hadoop.kafka.hive.KafkaStorageHandler'
tblproperties('oracle.kafka.table.key.type'='string',
'oracle.kafka.table.value.type'='long',
'oracle.kafka.bootstrap.servers'='nshgc0602:9092',
'oracle.kafka.table.topics'='topic1,topic2');
The example below shows the resulting Hive table. The Kafka key, value, offset, topic name, and partitionid are mapped to Hive columns. You can explicitly designate the offset for each topic/partition pair through a WHERE clause in you Hive query.
hive> describe test_table;
OK
topic string from deserializer
partitionid int from deserializer
key string from deserializer
value bigInt from deserializer
offset bigint from deserializer
timestamptype smallInt from deserializer
timestamp timestamp from deserializer
Time taken: 0.084 seconds, Fetched: 7 row(s) hive> SELECT * \
FROM test_table \
WHERE (topic="topic1" and partitoinid=0 and offset > 199) \
OR (topic="topic1" and partitionid=1 and offset > 198) \
OR (topic="topic2" and partitionid=0 and offset > 177) \
OR (topic="topic2" and partitionid=1 and offset > 176);hive> SELECT * FROM test_table WHERE timestamp > '2017-07-12 11:30:00'; See the Property Reference section below for descriptions of all table properties
Property Reference
Table 2-5 Table Properties of Hive Storage Handler for Kafka
| Property Name | Requirement | Description |
|---|---|---|
|
oracle.kafka.table.topics |
Required |
A comma-separated list of Kafka topics. Each Kafka topic name must consists of only letters (uppercase and lowercase), numbers, .(dot), _(underscore), and -(minus). The maximum length for each topic name is 249. These topics must have the same serialization mechanisms. The resulting Hive table consists of records from all the topics listed here. A Hive column “topic” will be added and it will be set to the topic name for each record. |
|
oracle.kafka.bootstrap.servers |
Required |
This property will be translated to the “bootstrap.servers” property for the underlying Kafka consumer. The consumer makes use of all servers, irrespective of which servers are specified here for bootstrapping. This list only impacts the initial hosts used to discover the full set of servers. This list should be in the form |
|
oracle.kafka.table.key.type |
Optional |
The key type for your record. If unset, then the key part of the Kafka record will be ignored in the Hive row. Only values of “string”, “integer”, “long”, “double”, “avro”, “avro_confluent”are supported. “string”, “integer”, “double” and “long” correspond to the built-in primitive serialization types supported by Kafka. If this property is one of these primitive types, then the Kafka key for each record will be mapped to one single Hive Column. If this property is set to “avro” or “avro_confluent”, then The difference between “avro” and “avro_confluent” is that the wire format for the serialization is slightly different. For “avro”, the entire bytes array of the key consists of the bytes of avro serialization. For “avro_confluent”, the bytes array consists of a magic byte, a version number, then the bytes of avro serialization of the key. |
|
oracle.kafka.table.value.type |
Optional |
The value type of your record. If unset, then the value part of Kafka record will be ignored in the Hive row. Use of this property is similar to use of |
|
oracle.kafka.table.key.writer.schema |
Optional |
An optional writer schema for the Kafka key’s Avro serialization. It’s required when the reader schema for the key is different from the schema in which the keys are written to Kafka brokers. It must be the exact schema in which Kafka keys are serialized. |
|
oracle.kafka.table.key.schema |
Required when “oracle.kafka.table.key.type” is “avro” or “avro_confluent” |
The JSON string for the Kafka key's Avro reader schema. It doesn't need to be exactly the same as the Kafka key's writer Avro schema. As long as the reader schema is compatible with the Kafka key or the converted object from the converter, it is valid. This enables you to rename Hive columns and choose what fields to keep from the Kafka key in the Hive row. If the schema in this property is different from the schema in which the Kafka keys are serialized, then |
|
oracle.kafka.table.value.writer.schema |
Optional |
An optional writer schema for the Kafka value’s Avro serialization. Its use is similar to |
|
oracle.kafka.table.value.schema |
Required when “oracle.kafka.table.value.type” is “avro” or “avro_confluent” |
The JSON string for the Kafka value's Avro reader schema. Its use is similar to |
|
oracle.kafka.table.extra.columns |
Optional, default to “true” |
A boolean flag to control whether to include extra Kafka columns: |
|
oracle.kafka.chop.partition |
Optional, default to false |
A Boolean flag to control whether to chop Kafka partitions into smaller chunks. This is useful when the number of Kafka partitions is small and the size of each Kafka partition is large. |
|
oracle.kafka.partition.chunk.size |
Optional |
When oracle.kafka.chop.partition is true, this property controls the number of Kafka records in each partition chunk. It should be set a value estimated by (Ideal size of a split)/(Average size of a Kafka record). For example, if the ideal size of a split is 256 MB and the average size of s Kafka record is 256 Bytes, then this property should be set to 1000000. |
2.1.6.2 Creating an Oracle Big Data SQL Table for Kafka Topics
Big Data SQL can use the ORACLE_HIVE access driver to query data stored in Hive Tables.
After you create a Hive table over Kafka data by using the Hive storage handler for Kafka, there are no special procedures for generating a Big Data SQL table from the resulting Hive table. The default ORACLE_HIVE settings can be overridden in the same way as with other Hive tables This is how to query the Hive external table that was created using the Hive storage handler for Kafka in the previous section.
CREATE TABLE test_table(
topic varchar2(50),
partitionid integer,
key varchar2(50),
value integer,
offset integer,
timestamptype integer,
timestamp timestamp
)
ORGANIZATION EXTERNAL
(TYPE ORACLE_HIVE DEFAULT DIRECTORY DEFAULT_DIR
ACCESS PARAMETERS
(
com.oracle.bigdata.cluster=hadoop1
com.oracle.bigdata.tablename=default.test_table
)
) PARALLEL 2 REJECT LIMIT UNLIMITED
Some Common Questions
-
Is Oracle Big Data SQL access to Kafka Brokers parallelized? For example, if I have six nodes running Oracle Big Data SQL, will all six nodes participate in a single query to Kafka so that we have to create a consumer group across all nodes to read the Kafka topic? Or, will only one node be used for a single SELECT statement to query from Kafka?
Like any Big Data SQL query, a query to Kafka will engage all of the nodes where Oracle Big Data SQL is installed.
-
In a Kafka query, how can we accommodate new incoming data? Can I have a query that waits (for a specified timeout) for new data to come into Kafka?
To pick up new data, you can run the Oracle Big Data SQL query periodically and filter by offset and timestamp to only retrieve the new rows (rows since the last read).
See Also:
The following section of the Big Data SQL Quick Start blog provides more information on accessing Kafka through Oracle Big Data SQL – Big Data SQL Quick Start. Big Data SQL over Kafka – Part 232.1.7 Creating an Oracle External Table for Object Store Access
The ORACLE_BIGDATA access driver enables you to create an external table over data stored in object stores. Oracle Big Data SQL currently supports access to Oracle Object Store and Amazon S3.
ORACLE_BIGDATA is primarily intended to support queries against object stores. It uses Smart Scan and Oracle Big Data SQL cells for scale out and performance against these stores. You can also use this driver to query local data, which is useful for testing and smaller data sets.
The ORACLE_BIGDATA driver is similar to the ORACLE_HDFS driver in that you create tables over the raw files. It does not use a metadata store like Hive. You specify the metadata as part of the table definition.
However, unlike ORACLE_HDFS, ORACLE_BIGDATA does not use Java drivers and the standard Hadoop mechanisms for data access (SerDes, InputFormats, etc.). ORACLE_BIGDATA uses optimized C-drivers for all data access. It supports the text, Avro, and Parquet file types. The text file type support is robust. For example, you can specify parameters for delimited text. You can also utilize Oracle's extensive JSON processing capabilities.
Steps for Accessing Data in Object Store
- Create a credential object
A credential object stores object store credentials in an encrypted format. The identity specified by the credential must have access to the underlying data in the object store.
- Create an Oracle Big Data SQL external table whose type is
ORACLE_BIGDATA. The create table statement must reference the credential object, which provides authentication against the object store. It also requires a LOCATION clause, which provides the URI to the files within the object store.
Creating the Credential Object
execute dbms_credential.create_credential(
credential_name => '<my_credential>',
username => '<username>',
password => '<password>'
);Then specify the credential object name in the com.oracle.bigdata.credential parameter.
See Also:
The Oracle Database PL/SQL Packages and Types Reference describes how to use DBMS_CREDENTIAL.CREATE_CREDENTIALDefining the LOCATION Clause
- For Oracle Cloud Infrastructure Object Storage, the URI format for files is:
location ('https://swiftobjectstorage.region.oraclecloud.com/v1/<object_storage_namespace>/<bucket>/<filename>') - For an Amazon S3 location, see https://docs.aws.amazon.com/AmazonS3/latest/dev/UsingBucket.html#access-bucket-intro. A possible example is:
location ('https://s3-us-west-2.amazonaws.com/adwc/<filename>')
If the URI is for a directory, then all files in the directory will be included.
If the credential parameter is omitted, then ORACLE_BIGDATA assumes LOCATION contains a file name and optional directory object name. In this case, the file resides in the local file system. The credential object is required for object store access only. If the directory object is not specified, the access driver uses the directory object specified in the DEFAULT DIRECTORY clause. The user that defined the table must have READ access to the directory object.
Setting Access Parameters
com.oracle.bigdata.fileformat={textfile|avro|parquet}
com.oracle.bigdata.credential.name=<credential object>
com.oracle.bigdata.credential.schema=<schema that owns the credential> For delimited text files, the rowformat parameter is also required.
See Also:
ORACLE_BIGDATA Access Parameters in this guide provides tables that list common access parameters as well as those specific to each file type.Note:
You can use ORACLE_BIGDATA to access local files for testing purposes or simple querying. In this case, the LOCATION field value is the same as what you would use for ORACLE_LOADER. You can use an Oracle directory object followed by the name of the file in the LOCATION field. For local files, a credential object is not required. However, you must have privileges over on the directory object in order to access the file.2.1.7.1 Create Table Example for Object Store Access
This section describes how to build a CREATE TABLE statement for object store access.
As shown in the example below, the required components are as follows:
-
The schema for the external table.
-
The credential object.
-
The fileformat parameter and any access parameters that are particular to the file format. For example, delimited Text files require the rowformat parameter. Parquet and Avro require only the fileformat parameter. Note that the default file format is Parquet.
-
The correct LOCATION clause syntax for the particular object store.
-
Use of a DEFAULT DIRECTORY clause with a LOCATION that is local rather than in an object store.
See Also:
ORACLE_BIGDATA Access Parameters for the full set of available parameters for each file type.Example: Accessing a File in an Object Store
Note:
Remember that for object store access you must first use DBMS_CREDENTIAL.CREATE_CREDENTIAL in the DBMS_CREDENTIAL PL/SQL package to create the credential object.exec dbms_credential.create_credential(
credential_name => '<my_credential_object_name>',
username => '<username>',
password => '<password>'
);Then within the ACCESS PARAMETER clause of the statement, assign the name of
the object to com.oracle.bigdata.credential.name as shown in the
statement. A Parquet file in an object store is the target in this example.
CREATE TABLE tkexbaseballtab
(date1 date,
date2 timestamp,
name varchar2(30),
nationality varchar2(20),
age number,
team varchar2(20),
active char(1),
average float,
payroll char(1),
exid VARCHAR2(20))
ORGANIZATION EXTERNAL
(TYPE ORACLE_BIGDATA
ACCESS PARAMETERS
(
com.oracle.bigdata.debug=TRUE
com.oracle.bigdata.credential.name=MY_CRED
com.oracle.bigdata.fileformat=parquet
)
location ('https://<domain>.com/BIGDATA_PARQUET/<filename>.parquet')
) REJECT LIMIT UNLIMITED;2.1.7.2 Accessing a Local File through an Oracle Directory Object
You can also use the ORACLE_BIGDATA driver to create an external table over a local file.
The statement below creates an external table over a text file in a local
directory. For a JSON file, set the file format to textfile. No other access
parameters are needed. You do not need to supply a credential when accessing local files. Do
the same for a delimited text file, but if the fields terminator and/or line terminator used
in the file are other than the default values, define those as well. In this example, we set
the field terminator to the vertical bar (‘|’). This statement does not
include a DEFAULT DIRECTORY clause. Instead in the LOCATION clause we include the name of the
directory object where the file is located – DEF_DIR1 .
CREATE TABLE b19724081
( CONTRACT_STATUS_CODE VARCHAR2(1),
CONTRACT_STATUS_DESCRIPTION VARCHAR2(200),
POSTING_AGENT VARCHAR2(50),
DATA_ORIGIN VARCHAR2(50),
WAREHOUSE_POSTING_TIMESTAMP DATE,
WAREHOUSE_UPDATE_TIMESTAMP DATE
)
ORGANIZATION EXTERNAL
( TYPE ORACLE_BIGDATA
DEFAULT DIRECTORY "DEF_DIR1"
ACCESS PARAMETERS
(
com.oracle.bigdata.fileformat=textfile
com.oracle.bigdata.csv.rowformat.fields.terminator = '|'
)
location ("DEF_DIR1":"<'filename>.csv')
)
REJECT LIMIT UNLIMITED;Compression Format
There is no compression format support for textfile file format. In other words, the files for textfile must be uncompressed. We support gzip, snappy, lzo, and zlib for Parquet and Avro file format. The compression format is detected automatically, and there is no parameter needed.
2.1.7.3 Parquet to Oracle Data Type Conversions
Oracle Big Data SQL support conversions from Parquet file data types to Oracle data types, but for scalar fields only.
Note:
Any non-supported conversions cause the SELECT from the external table to fail. Parquet does not require that every file have data/metadata for every field. In that case, the unsupported conversion will not be detected until the metadata for that field is read later.
If data for a column encounters a conversion error, for example, the target column is not large enough to hold the converted value, the value for the column will be set to NULL.
| Parquet File Data Types | Supported Conversions to Oracle Data Types |
|---|---|
|
FLOAT, DOUBLE, DECIMAL |
NUMBER, BINARY_FLOAT, BINARY_DOUBLE, CHAR, NCHAR, VARCHAR2, NVARCHAR2, CLOB, NCLOB |
|
DATE |
NUMBER – number of days BINARY_FLOAT, BINARY_DOUBLE – number of days CHAR, NCHAR, VARCHAR2, NVARCHAR2, CLOB, NCLOB – string for the represented date using NLS parameters for the session DATE, TIMESTAMP |
|
TIME_MILLIS, TIME_MICROS |
NUMBER – number of milliseconds or microseconds BINARY_FLOAT, BINARY_DOUBLE – number of milliseconds or microseconds CHAR, NCHAR, VARCHAR2, NVARCHAR2, CLOB, NCLOB – string for the represented time using NLS parameters for the session INTERVAL DAY TO SECOND |
|
TIMESTAMP_MILLIS, TIMESTAMP_MICROS, TIMESTAMP(from impala) |
CHAR, NCHAR, VARCHAR2, NVARCHAR2, CLOB, NCLOB – string for the represented timestamp using NLS parameters for the session DATE, TIMESTAMP |
|
UTF8, ENUM |
NUMBER BINARY_FLOAT, BINARY_DOUBLE CHAR, NCHAR, VARCHAR2, NVARCHAR2, CLOB, NCLOB DATE, TIMESTAMP RAW, BLOB (interpret as hex string) INTERVAL |
|
JSON |
CHAR, NCHAR, VARCHAR2, NVARCHAR2, CLOB, NCLOB |
2.1.7.4 Avro to Oracle Data Type Conversions
Support for Avro to Oracle data type conversions is shown in the table below.
Table 2-6 ORACLE_BIGDATA Mappings From Avro to Oracle Data Types
| Avro | ORACLE_BIGDATA Mapping |
|---|---|
| STRING | VARCHAR2 |
| BYTES | BLOB |
| INT | NUMBER(10) |
| LONG | NUMBER(19) |
| FLOAT | BINARY_FLOAT |
| DOUBLE | BINARY_DOUBLE |
| BOOLEAN | NUMBER(1) |
| NULL | VARCHAR2(1) BYTE |
| ENUM | VARCHAR2 |
| UNION | BINARY_DOUBLE |
| FIXED | BLOB |
| DECIMAL(p) | NUMBER(p) |
| DATE | DATE |
| DURATION | BLOB |
| TIME-MILLIS | VARCHAR2(20 BYTE) |
| TIME-MICROS | VARCHAR2(20 BYTE) |
| TIMESTAMP_MILLIS | TIMESTAMP(3) |
| TIMESTAMP_MICROS | TIMESTAMP(6) |
| RECORD | CHAR, VARCHAR, or CLOB |
| ARRAY | CHAR, VARCHAR, or CLOB |
| MAP | CHAR, VARCHAR, or CLOB |
2.1.7.5 ORACLE_BIGDATA Support for Compressed Files
ORACLE_BIGDATA driver support for access to compressed files in object stores is as follows:
- Compressed Text files
Not supported at this time. Text files must be uncompressed.
- Compressed Parquet and Avro files
The driver can read from Parquet and Avro files compressed with gzip, snappy, lzo, or zlib. The compression format is detected automatically. No parameters to handle compressed files are needed in the external table create statement.
2.2 Querying External Tables
Users can query external tables using the SQL SELECT statement, the same as they query any other table.
Note:
The MODIFY EXTERNAL clause is not allowed for any external table created through the ORACLE_BIGDATA driver.2.2.1 Granting User Access
Users who query the data on a Hadoop cluster must have READ access in Oracle Database to the external table and to the database directory object that points to the cluster directory. See "About the Cluster Directory".
2.2.2 About Error Handling
By default, a query returns no data if an error occurs while the value of a column is calculated. Processing continues after most errors, particularly those thrown while the column values are calculated.
Use the com.oracle.bigdata.erroropt and com.oracle.bigdata.overflow parameters to determine how errors are handled.
2.2.4 About File Readers
2.2.4.1 Using the Custom Parquet Reader for Oracle Big Data SQL
For reading parquet files, you have the option of using the custom Parquet reader for Oracle Big Data SQL. This proprietary driver improves performance and makes more efficient use of cluster resources.
Disabling or Re-Enabling the Custom Parquet Reader
com.oracle.bigdata.useOracleParquet=falseCompatibility with Previously Created Parquet Format Data
Use of the customer reader requires no changes to data format. However, for best performance, the format must provide min and max values for each column for each Parquet block. These values are used by the standard Hadoop Parquet InputFormat, as well as the custom Parquet reader, to optimize the query. The resulting optimization significantly improves query performance with both Hive and Oracle Big Data SQL.
Note that Parquet files created by Impala do not include min and max values for each column for each Parquet block.
To ensure that min and max values are available, it is recommended that you write Parquet files with Hive or other tools that generate output in the standard Hadoop Parquet InputFormat, such as PrestoDB and Spark.
To check if a file includes these values, you can use the parquet tools JAR to dump information about the file:
# hadoop jar parquet-tools-1.5.0-cdh5.12.0.jar meta <filename>.parqOn a CDH Hadoop distro, the parquet-tools command may also be configured in your path.
2.3 About Oracle Big Data SQL on the Database Server (Oracle Exadata Machine or Other)
This section explains the changes that the Oracle Big Data SQL installation makes to the Oracle Database system (which may or may not be an Oracle Exadata Machine).
2.3.1 About the bigdata_config Directory
The directory bigdata_config contains configuration information that is common to all Hadoop clusters. This directory is located on the Oracle Database system under $ORACLE_HOME/bigdatasql. The oracle file system user (or whichever user owns the Oracle Database instance) owns bigdata_config . The Oracle Database directory ORACLE_BIGDATA_CONFIG points to bigdata_config.
2.3.2 Common Configuration Properties
The installation store these files in the bigdata_config directory under $ORACLE_HOME/bigdatasql :
The Oracle DBA can edit these configuration files as necessary.
2.3.2.1 bigdata.properties
The bigdata.properties file in the common directory contains property-value pairs that define the Java class paths and native library paths required for accessing data in HDFS.
These properties must be set:
The following list describes all properties permitted in bigdata.properties.
bigdata.properties
| Property | Description |
|---|---|
|
bigdata.cluster.default |
The name of the default Hadoop cluster. The access driver uses this name when the access parameters do not specify a cluster. Required. Changing the default cluster name might break external tables that were created previously without an explicit cluster name. |
|
bigdata.cluster.list |
A comma-separated list of Hadoop cluster names. Optional. |
|
java.classpath.hadoop |
The Hadoop class path. Required. |
|
java.classpath.hive |
The Hive class path. Required. |
|
java.classpath.oracle |
The path to the Oracle JXAD Java JAR file. Required. |
|
java.classpath.user |
The path to user JAR files. Optional. |
|
java.libjvm.file |
The full file path to the JVM shared library (such as |
|
java.options |
A comma-separated list of options to pass to the JVM. Optional. This example sets the maximum heap size to 2 GB, and verbose logging for Java Native Interface (JNI) calls: Xmx2048m,-verbose=jni |
|
java.options2 |
A space-delimited list of options to pass to the JVM. Optional. The delimiter must be a space character, not a tab or other whitespace character. This example sets the maximum heap size to 2 GB, and verbose logging for Java Native Interface (JNI) calls: Xmx2048m -verbose=jni Note: Notice thatjava.options is comma-delimited, while java.options2 is space delimited. These two properties can coexist in the same bigdata.properties file.
|
|
LD_LIBRARY_PATH |
A colon separated (:) list of directory paths to search for the Hadoop native libraries. Recommended. If you set this option, then do not set java.library path in |
Example 2-1 shows a sample bigdata.properties file.
Example 2-1 Sample bigdata.properties File
# bigdata.properties # # Copyright (c) 2014, Oracle and/or its affiliates. All rights reserved. # # NAME # bigdata.properties - Big Data Properties File # # DESCRIPTION # Properties file containing parameters for allowing access to Big Data # Fixed value properties can be added here # java.libjvm.file=$ORACLE_HOME/jdk/jre/lib/amd64/server/libjvm.so java.classpath.oracle=$ORACLE_HOME/hadoopcore/jlib/*:$ORACLE_HOME/hadoop/jlib/hver-2/*:$ORACLE_HOME/dbjava/lib/* java.classpath.hadoop=$HADOOP_HOME/*:$HADOOP_HOME/lib/* java.classpath.hive=$HIVE_HOME/lib/* LD_LIBRARY_PATH=$ORACLE_HOME/jdk/jre/lib bigdata.cluster.default=hadoop_cl_1
2.3.2.2 bigdata-log4j.properties
The bigdata-log4j.properties file in the common directory defines the logging behavior of queries against external tables in the Java code. Any log4j properties are allowed in this file.
Example 2-2 shows a sample bigdata-log4j.properties file with the relevant log4j properties.
Example 2-2 Sample bigdata-log4j.properties File
# bigdata-log4j.properties
#
# Copyright (c) 2014, Oracle and/or its affiliates. All rights reserved.
#
# NAME
# bigdata-log4j.properties - Big Data Logging Properties File
#
# DESCRIPTION
# Properties file containing logging parameters for Big Data
# Fixed value properties can be added here
bigsql.rootlogger=INFO,console
log4j.rootlogger=DEBUG, file
log4j.appender.console=org.apache.log4j.ConsoleAppender
log4j.appender.console.target=System.err
log4j.appender.console.layout=org.apache.log4j.PatternLayout
log4j.appender.console.layout.ConversionPattern=%d{yy/MM/dd HH:mm:ss} %p %c{2}: %m%n
log4j.appender.file=org.apache.log4j.RollingFileAppender
log4j.appender.file.layout=org.apache.log4j.PatternLayout
log4j.appender.file.layout.ConversionPattern=%d{yy/MM/dd HH:mm:ss} %p %c{2}: %m%n
log4j.logger.oracle.hadoop.sql=ALL, file
bigsql.log.dir=.
bigsql.log.file=bigsql.log
log4j.appender.file.File=$ORACLE_HOME/bigdatalogs/bigdata-log4j.log2.3.3 About the Cluster Directory
The cluster directory contains configuration information for a Hadoop cluster. Each cluster that Oracle Database accesses using Oracle Big Data SQL has a cluster directory. This directory is located on the Oracle Database system under $ORACLE_HOME/bigdatasql/clusters/. For example, cluster bda1_cl_1 would have a directory $ORACLE_HOME/bigdatasql/clusters/bda1_c1_1 and under $ORACLE_HOME/bigdatasql/clusters/bda1_c1_1/config would be the following files for client configuration files for accessing the cluster:
-
bigdata.hosts(not editable by customers) -
core-site.xml -
hdfs-site.xml -
hive-site.xml -
mapred-site.xml(optional) -
log4jproperty files (such ashive-log4j.properties)
$ORACLE_HOME/bigdatasql/databases/<database name>/bigdata_config/default_cluster is a soft link to the directory of the default cluster.
A database directory object points to the cluster directory. Users who want to access the data in a cluster must have read access to the directory object.
2.3.4 About Permissions
On the Oracle database server, ensure that the oracle user (or whatever user owns the Oracle Database installation directory) has READ/WRITE access to the database directory that points to the log directory.
On the Hadoop side, when you run Database Acknowledge (# ./jaguar databaseack [config file]) this operation creates an account for the database owner and grants required permissions.