Wednesday, 7 September 2016

34 Powerful Complex Teradata Architectural Questions and Answers

Teradata Architectural Questions
Teradata Architectural Questions 
QUESTION No:1
An administrator is trying to manage resources on the system because there is a business critical reporting requirement. The reporting workload starts at 9AM and is required to finish by 10AM. At the same time, ad hoc users are allowed to access the system. The administrator is building a Query Resource Throttle in TDWM to better manage the resources. Which two are available as options for building the throttle? (Choose two.)

A. Index Scan Types Allowed
B. Throttle All-AMP Queries Only
C. No Aggregate Functions Allowed
D. Only limit queries with Step Time Threshold

Answer: C,D

QUESTION NO: 2
An administrator is asked to keep a daily backup of a very large transaction table. The table is defined with a PPI with a daily partitioning expression on insert date and the table is insert only. There is not enough of a window to backup the entire table each night. In this situation, which backup strategy is the best approach?

A. Implement a cluster backup of the table.
B. Implement an online archive of the table.
C. Implement an all-AMP backup of the table.
D. Implement partition backup for the previous day.

Answer: C

QUESTION NO: 3
Which two statements are true about FastLoad capabilities regarding error handling, return codes, and handling statistics about the load job? (Choose two.)

A. The predefined system variables, prefixed with &SYS, are not implemented in FastLoad.
B. Notify Exit routines specify a predefined action to be performed whenever a failure occurs in the utility. The action only handles return codes.
C. Predefined system variables, prefixed with &SYS, are used to capture load statistics and can be used to determine if the process ran successfully.
D. Notify Exit routines specify a predefined action to be performed whenever a failure occurs in the utility. The action can handle return codes and load statistics.
Answer: B,C

QUESTION NO: 4

A customer has a requirement to load data from a single message queue into a Sales table that is part of an active data warehouse. The Sales table has no additional indexes. The data latency requirements are daily. What is the most effective approach to support this requirement?

A. Use TPump to continuously read from the queue and insert the data.
B. Pull the data from the queue into a file and use TPump to insert the data from the file.
C. Pull the data from the queue into a file and use MultiLoad to insert the data from the file.
D. Pull the data from the queue into a file and use FastLoad to load the data into a staging table and then use SQL INSERT SELECT.

Answer: B

QUESTION NO: 5
Which type of physical device, in addition to Enterprise System Connection (ESCON), is required to allow a mainframe in one city to connect to a Teradata Database in another city?

A. ESCON Director
B. Block Multiplexer
C. Channel Extender
D. BYNET Optical Extension (BYOX)

Answer: A

QUESTION NO: 6
Which method can be used to address the privacy requirements of an application that accesses personal information?

A. Use Access Logging to determine the requesting user, type of access, and the referenced object.
B. Extract the data, define a masking algorithm, and insert the data, with masking, into the response.
C. Use multi-factor authentication to provide strong access control and enhance the safety of data stored in the database.
D. Create a user defined function that can only be accessed from within the application to encrypt and decrypt the personal data.

Answer: A

QUESTION NO: 7
An administrator needs to establish a data warehouse architecture for a user community of 2000 end-users and a data integration group of 20 users. Which three items are important when creating the database hierarchy to support these two groups of users? (Choose three.)

A. priorities
B. roles and profiles
C. implicit access rights
D. access and DBQL logging
E. permanent space allocation
F. archive and restore operations

Answer: A,D,F

QUESTION NO: 8
A programmer has executed a query and wants to know performance statistics. Upon investigating DBQL the query is not found. The programmer wants to see the query. Which two administrative options are available for making DBQL available sooner? (Choose two.)

A. End logging on any user will flush the cache.
B. Reduce the setting of the dbscontrol parameter dbscachectrl.
C. Reduce the setting of the dbscontrol parameter dbqlflushrate.
D. End logging on systemfe is the only user that will allow the cache to be flushed manually.

Answer: B,C

QUESTION NO: 9
A new workload management practice is being established without the involvement of application or user communities. What information should the administration teams collect to establish the workload management policy?

A. AMPUsage
B. DBQL.QryLog
C. ResUsageSpma
D. Metadata services
E. Dictionary access counts

Answer: A

QUESTION NO: 10
A critical customer identification application is used by all end users. Several sets of tactical queries have been created and deployed which select and transform the application data to support day-to-day credit extension, account management, and offer delivery decisions. Over time, 15 single and multi-table join indexes have been created on the critical tables to attempt to cover these tactical queries. A new set of tactical queries is being developed to support a mandated fraud reporting initiative required by the government. The application development team recommends adding another two multi-table join indexes to the application tables. The ETL team is consistently missing the required load window for this data which has resulted in financial penalties to the company for unmet Service Level Agreements (SLAs). The database administrator has been asked to review the database architecture and recommend potential solutions.

Which statement is true about using join indexes in the existing database architecture?
A. The new set of tactical queries might use more than one of the existing join indexes and therefore not require the new join indexes.
B. There is a limit of 16 join indexes per table so the addition of the two new join indexes may necessitate the removal of at least one of the existing join indexes.
C. The number of times a query (for which a join index was designed) must be executed to offset the creation and maintenance time of the join index should be calculated.
D. As new rows are added to the base tables, multiple join indexes can be updated as part of the same transaction so the incremental cost of adding new join indexes is mitigated.
Answer: A

>> Page two Questions

No comments:

Post a Comment