1. What is a view and how to create one?

A view is simply the representation of a SQL statement that is stored in memory so that it can easily be re-used. For example, if you frequently use a query SELECT empid FROM emp, you may create a view as
CREATE VIEW view_emp
AS
SELECT empid FROM emp;

Then every time you query this view, Oracle will go out and execute the SQL behind the view to query the database data.

2. Can Oracle view improve performance? Why or why not?

In most cases view does improve performance because of commonality of code being used. Since a view is based on one common set of SQL, when it is called it’s less likely to require parsing. This is because the basic underlying SQL that is called is always the same. A query without parsing is faster. However, since you can add additional where clauses when calling a view, you still need to use bind variables. Additional where clauses without a bind variable can still cause a hard parse.

3. Explain how you may use view to enhance security and what are view’s limitations for security practice?

It is a two-question in one but people often miss the second one. In any database, view’s limitation for security is the issue of scalability.

Answer:

Views are commonly used to hide the tables that actually contain the data you are querying. Also, views can be used to restrict the columns that a given user has access to. Using views for security on less complex databases is probably not a bad thing. As databases become more complex, this solution becomes harder to scale and other solutions will be needed.

4. In what situation Oracle views can affect performance, and what is Oracle’s solution to this issue?

When you create complex views with many table joins to represent large subsets of query and allow many developers and end users to access these views, you may experience poor performance. In this case, you may consider Oracle’s materialized view.

5. What is materialized view and where is it used?

A materialized view is a snapshot of data that can be refreshed based on a schedule. It is a replica of a target master from a single point in time. Oracle uses materialized views to replicate data to non-master sites in a replication environment and to cache expensive queries in a data warehouse environment.

6. Can view be indexed in Oracle?

Don’t think Oracle can do everything and quickly answer ‘yes’. SQL Server offers view index, but Oracle does not do it.

Answer:

Regular Oracle view cannot be indexed, but you can utilize table level indexed to tune underlying query that generates the view.

7. Can you insert and update records in a view?

This question asks if you can execute insert or update statement directly on a view instead of on a table.

Answer:

You can only perform data manipulation language (DML) operations with simple views. Complex views don’t support DML. view is mainly used for presenting and data and simplifying query. Data insert and update should be done on table level.

8. What’s the different between GRANT CREATE VIEW and CREATE ANY VIEW?

It is not a question about view. This question checks your knowledge of controlling user privileges in Oracle database. Grant Create view gives user’s right to create new views in grantee’s schema only. GRANT CREATE ANY VIEW allow users to create views in any schema.

“I have extensive DBA experience in configuring database server for both OLTP and OLAP systems.  I worked on many transaction-oriented database, I also had the opportunity to setup database for decision support systems.  Nowadays most of the RDBMS provide options to setup OLAP for you during installation, you simply select this option and the installation process will configure your database suitable for OLAP and DDS applications.  Of course I can always change my database configuration manually when necessary.  The major difference of OLAP comparing to transaction database is that OLAP requires fast data retrieval and concurrent access.  Many DSS applications run very complicated queries.  In general, there are every few data manipulation activities such as insert, update, and delete.  Therefore in OLAP configuration I use indexes heavily to speed up data query process.  Oracle’s bitmap index and materialized views are also helpful.  I would also configure table partitions based on the analysis of data structure.  I would design tablespaces on different data files put on different hard disks and build tables across. That would increase concurrent access and reduce I/O bottleneck.  For DSS system, I may consider de-normalize the data structure and build dement ional database for reporting and BI purpose.  Depending on the specific project, I may implement other options such as configuring the database cache to meet the OLAP and DSS needs.”

“In my past job I’ve configured SQL Trace and Autotrace in our Oracle production environment with both SQLPlus and Toad for Oracle tools. I used both SQL Trace and Autotrace almost daily for SQL query analysis and performance tuning. I created explain plan tables for all users who might use Autotrace in SQLPlus. For users who use Toad, I only needed to create one set of explain plan tables under Toad schema and the plan tables can be shared by multiple Toad users. To configure SQL Trace, I need to alter session to set it up. Then I turn the trace on before query execution and turn it off after the SQL run. If I run it in SQLPlus, I would need to use the database utility TKPROF to convert the trace output into readable file with formatted data. In Toad I could the GUI interface to go through the SQL trace process.”

The follow-up question could ask you what is the difference between SQL Trace and Autotrace, and how you determine which one to use first. So your answer can further explain the knowledge like:

“Autotrace provides us with explain plan and performance statistics on the actual execution of the SQL run. SQL Trace provides performance statistics on the SQL run. It must be used in conjunction with the Database utility TKPROF in order to provide meaningful information. The performance statistics for the SQL Trace our more configurable and provide more information; however, Autotrace is easier to get feedback immediately. Often, Autotrace is used first then SQL Trace when more information is needed.”

As for database design tools, there are many.  The most popular ones are:
Rational Rose (from IBM)
ERWin (from CA)
Oracle Designer (from Oracle)
PowerDesigner (from Sybase)

A typical answer could be:

“I mainly used ERWin for database design in my previous projects.  I used this tool to collect business entity data and did data modeling with entity relationship diagram, and then I used the tool’s function to build the logical database based on the ERD.  I also use it to help me in physical database design analysis and generate SQL scripts to build database objects such as tables, indexes, and views.  I’ve used ERWin for 4 years in Oracle and DB2 database design projects.  But when I worked on Microsoft SQL Server, I used the build-in Enterprise Manager to do data modeling and create ERD.”

“I have hands-on skills and experience in database design and data modeling.  In previous jobs I have worked as the technical lead to design large scale database systems in Oracle, Sybase, and Microsoft SQL Server.  I have experience in several database design tools such as Oracle Designer, PowerDesigner, and ERWin.  Using these tools I collected and documented user requirements, analyzed business rules and entities, created entity relationship diagrams, designed logical database based on the ERD, and generated SQL scripts to build the physical database., and configured the database server and setup the database.  I also used database design tools to reverse-engineer existing database when necessary.  Database design tools are very helpful for me to automate many tasks through the database design lifecycle, but in many cases I still need to fine-tune the design manually.  I also have experience in using design tools to create data repository for meta data management.  In one of my jobs I worked as the repository administrator responsible for building and maintaining Oracle data repository in Oracle Designer for our term members to work together in a common data design platform through database design phases.  ”

Suggested answer:

“I have 6 years of database design experience mainly with Oracle database and have been involved in over 10 database design projects. In these projects I’ve worked on analysis of business entities and data modeling, creating entity relationship diagrams, design the logical database such as converting entities and attributes into table and columns. I also worked a lot on designing the physical database such as analyzing storage usage and performance, designing tablespace and table partition, writing SQL scripts to create database objects like tables, views, and indexes. One of the databases I designed is for a Federal agency’s employee activity tracking system; it contains over 200 tables, about 50 views, and more than 150 indexes on an Oracle server in UNIX platform. I used Oracle Designer to assist most of my database design tasks.”

“In most of data migration projects I developed customized scripting processes.  I’ve developed BCP scripts along with Windows’ batch scripting to migrate data from flat files into Sybase database.  I used Oracle’s SQL Loader a lot to load ASCII data into Oracle server.  I also had a chance to user Siebel’s Enterprise Integration Manager (EIM) to convert data from external source into Siebel core database structure.”

In general, database development experience is focused on programming on the database server, – writing store procedures, triggers, and packages.  In Oracle, you use PL/SQL; in Microsoft SQL Server and Sybase, you use T-SQL.  If you can find out what database server they are using, we then can emphases your skill set on the corresponding SQL language.  The following sample answer is to target Oracle database development:

Sample answer:

“I have over 5 years of experience in SQL and PL/SQL programming on Oracle server.  In my last project I was the lead developer for database programming.  This project is a client/server application for inventory tracking.  All business rules were processed on the backend database server.  I developed more than 50 Oracle stored procedures and triggers.  Many stored procedures were organized into packages.  Those procedures were used to process complicated rules and automate inventory tracking in our asset management system.  In previous projects I also worked on Microsoft SQL Server database and coded many stored procedures and functions in T-SQL.”

These are fundamental questions to test your knowledge of data warehousing. To answer the questions, just briefly explain each phase of the lifecycle:

1. Data acquisition processes—extract, transform, load (ETL)
The processes include sourcing, cleansing, transforming, and aggregating data using parallel technology tools to build industrial-strength ETL processes that accommodate high data volumes from disparate sources. We identify the best sources for data elements, reconstructing data when required and deploying the most appropriate tools to retrieve the data from its primary sources. Through the cleansing process, we enhance data quality by ensuring data accuracy, type, and consistency, as well as eliminating duplicate records.

2. Data repositories
In the phase you need to build a variety of data repositories, including operational data stores, data marts, data warehouses, web warehouses, and data hubs. You start by implementing and properly documenting a physical data model, ensuring data from all functional areas is sufficiently integrated to support cross-functional analyses. Then you perform database tuning, model denormalization, and aggregation as necessary to support information delivery requirements. When scalability requirements call for it, you partition and distribute data into a parallel architecture.

3. Information delivery applications
You implement information delivery applications that allow corporate users to access the data in the warehouse. These include decision support tools, data mining and analytic tools, and applications that optimize supply chain, campaign management, billing, and industry-specific processes. In this phase you also ensure near- and long-term reporting and access requirements are met as well. These may include fixed-frequency static reports; ad-hoc reports; dynamic, multidimensional queries; Internet/intranet application interfaces; and data mining.

4. Data warehouse administration
As the data warehouse grows, administration (or management) of the repository is a crucial step in optimizing results and return on investment. You provide data warehousing administration services such as performance analysis, user analysis, benchmarking, auditing, and tuning to help clients measure the ongoing success of their data strategies.