Enabling Faster Data Loads from Teradata Data Sources

In This Section:

Using Teradata Data Sources

Installing Required Teradata Software

Setting Up the Environment for Using Export Operator

Loading Teradata Data Using Teradata Parallel Transporter

Support for Unicode and Multibyte Character Sets

Using Teradata Data Sources

You can use Teradata Parallel Transporter (TPT) from Teradata Tools and Utilities to significantly improve data load performance. With this method, ODBC is used to extract the database schema; then TPT retrieves the data.

Essbase supports Teradata Parallel Transporter (TPT) 12 with the Teradata 12 ODBC driver on the following platforms: 32-bit and 64-bit AIX, 32-bit HP-UX, Red Hat Linux, 32-bit and 64-bit Solaris, 32-bit Windows, and 64-bit Windows AMD. TPT 12 can connect to Teradata V2R6.x and V12.x databases, not to Teradata V2R5.x databases. Consult your Teradata documentation for supported configurations.

For information about the versions of Teradata databases that Essbase supports as data sources, and supported Teradata ODBC drivers, see the Oracle Hyperion Enterprise Performance Management System Installation Start Here.

Installing Required Teradata Software

The customer is responsible for having the correct Teradata license and ODBC version installed and configured on the Essbase Server computer. See the Teradata documentation for installation instructions.

  • From Teradata Tools and Utilities, install Teradata Parallel Transporter Export Operator, Shared ICU Libraries for Teradata, Teradata GSS Client, and CLI. (For Linux installations, select libraries built by GCC 3.3.)

  • Install the Teradata ODBC driver.

Setting Up the Environment for Using Export Operator

Follow the instructions in Configuring Data Sources, and then perform these tasks:

  • Add an entry to the hosts file for the Teradata database; for example:

    172.27.24.181      tera2db     tera2cop1
  • Configure a system ODBC DSN for $TELAPI$<tera> where <tera> is the name of the Teradata data source: for example:

    DSN = $TELAPI$tera2db
  • For UNIX operating systems, ensure needed environment variable paths are defined in the appropriate location (the Windows installation automatically updates needed environment variables):

    • TD ODBC driver

    • CLIv2

    • TD GSS

    • Shared ICU

    • TPT export operator files

    • DataDirect ODBC driver

  • In addition, in the appropriate path for the operating system, set the following variables for Teradata Parallel Transporter. (For details, see the “Code Samples” appendix in Teradata Parallel Transporter Application Programming Interface Programmer Guide); for example, for Solaris SPARC:

    • export LD_LIBRARY_PATH = <library path>:$LD_LIBRARY_PATH

      export LD_LIBRARY_PATH = /usr/tbuild/12.00.00/lib:$LD_LIBRARY_PATH
    • export NLSPATH = <directory path of the catalog>/%N:$NLSPATH

      export NLSPATH = /usr/tbuild/12.00.00/msg/%N:$NLSPATH
    • (If CLI is not installed in the default directory) export COPERR = <directory location of errmsg.cat >

      export COPERR = /usr/lib

Loading Teradata Data Using Teradata Parallel Transporter

Follow the instructions in Loading SQL Data. When you open the SQL data source, select the desired data source name with the prefix $TELAPI$ that you defined as the ODBC DSN. For the SQL statement, define a native Teradata query in the SQL SELECT, FROM, and WHERE statements. Do NOT include carriage returns or line feeds in these statements. Each entry must be in a single statement. See the relevant Teradata documentation for native Teradata SQL query rules.

Support for Unicode and Multibyte Character Sets

Teradata supports multibyte character set (MBCS) and Unicode text, which Essbaseretrieves using TPTapi.

To use this functionality, perform these tasks:

  • Verify that the client character set that Essbase uses in installed or enabled in the Teradata database.

  • Make sure that the character set of the ODBC driver matches the character set that Oracle Essbase passes to TPTapi.

    To do so, you should create the ODBC connection DSN with the character set name that matches that used by the $ESSLANG variable, as shown in Table 2.

    Table 2. Supported Character Sets

    Character Set$ESSLANG Variable Setting
    Latin (covers almost all western languages)Various
    JapaneseKANJISJIS_05
    UnicodeUTF8

Note:

Essbase retrieves data in the supported character set; however, the SQL queries must be in English.