Sie sind auf Seite 1von 3

ding interviews on netezza technology and getting selected in the interview is v ery easy.

Just go through the below interview questions and definitely you will get selected in the interview. 1. What are the data warehouse appliances you know? Netezza Greenplum Vertica Teradata Dataupia Oracle Exadata 2. What are the environment variables that are required to connect to netezza? The environment variables required are: NZ_HOST, NZ_DATABASE, NZ_USER, NZ_PASSWO RD 3. What are the different states of Netezza? Online: Normal or usual state. Stopped: Netezza will shut down after completing current queries, no new queries allowed. Offline: Waits for completion of current queries, new queries in queue receives error. Paused: Same as above, but no error displayed. Typically caused during Netezza b oot up or startup. Down: Just plain down, could be due to Netezza server problem or user initiated. 4. What are the constraints on a table are enforced? The only constraint netezza supports is Not null. There are no primary key and f oreign key constraints in netezza. 5. Can you insert duplicate rows in netezza table? Yes. As there are no primary key constraints in netezza you can insert duplicate rows. 6. How the NOT NULL specification on a column improves the netezza performance? Specifying Not Null on each column in table results in better performance. Netez za tracks the NULL values at rowheader level. Having NULL values results in stor ing references to NULL values in header. If all columns are NOT NULL, then there is no record header. 7. How FPGA can be helpful in improving query performance? While reading data from the disk, the Field Programmable Gate Array (FPGA) on ea ch SPU filters out unwanted data. This process of data elimination removes IO bo ttlenecks and frees up downstream components such as the CPU, memory and network from processing extra data. 8. What is a snippet? A snippet is a small unit of work that are carried out in SPU. 9.What are zonemaps? An extent is the smallest unit of disk allocation on a SPU. Zonemaps is internal mapping structures to the extents that take advantage of the internal ordering of data to eliminate extents that do not need to be scanned. Zonemaps transparen tly avoid scanning of unreferenced rows. Zonemaps are created for every column i n the table and contain the minimum and maximum values for every extent. 10. How the zonemaps are created and updated? Zonemaps are created and refreshed for every SPU when you Generate statistics, N zload operation, Insert, update operations, Nzreclaim operation.

11. What is generate statistics and generate express statistics OR what is the d ifference between generate statistics and generate express statistics? Generate statistics is used to gather statistics about each table column's propo rtion of duplicate values, minimum values, maximum values, null values, unique v alues and updates the system catalog tables. The difference between 'generate statistics' and 'generate express statistics' i s based on how the column uniqueness is calculates. The 'generate express statis tics' calculates estimated dispersion values based on the sampling of rows in th e table. 'Generate express statistics' uses approximation in generating the stat s where as 'generate statistics' uses all the rows in the table. 12. What is the use of creating materialized views? A materialized view reduces the width (number of columns) of data being scanned in the base table by creating a thin version (fewer columns) of the base table t hat contains a small subset of frequently queried columns. 13. What is the distribution of materialized views? A materialized view has the same distribution key as the base table. 14. What are the limitations of materialized views? You cannot insert, update, delete or truncate a materialized view. Any changes o n the base tables will reflect into materialized views. You can specify only one base table in the from clause. Base table can't be a external table, system table or a temporary table. You cannot use a where clause in the materialized view. Expressions are not allowed as columns. Why integer data type is preferred in Netezza. A couple of reasons: 1) Better joins, thus effecient. 2) Netezza compress works only for integer type of data, not with varchar or dat e. 3) Zonemaps are based on integer data types. 15. What are the best practices of creating materialized views? Create materialized views with few columns which are frequently queried. Specify order by clause on the most restrictive columns (columns used in where c lause). Periodically or manually refresh the materialized views. 1. What are the partitioning methods available in netezza? There are two partitioning methods available in netezza: Random partitioning: Distributes the data randomly. Hash Partitioning: Distributes the data on the specified columns. 2. Up to how many columns you can specify in distribute on clause? You can specify up to four columns in the distribution clause? 3.If you did not specify any distribute on clause while creating a table, what d istribution netezza uses? Netezza distributes the data on the first column and it uses Hash partitioning 4. Can you update the columns used in distribution clause? No, the column that is used in distribution clause cannot be used for updates. 5. What data types are most suited for the columns specified in distribution cla use? Integer

6. How do you redistribute a table? Use Create Table As (CTAS) to redistribute the data in a table. While creating t he new table specify the distribute on clause to distribute the data on the new columns. 7. If you did not specify any distribution clause, how the Create Table AS (CTAS ) will distribute the rows? CTAS will get distribution from the original table. 8.How do you check the rows in a table are equally distributed in all SPU's or n ot? To check the distribution of rows run the following query SELECT datasliceid, COUNT(*) FROM <table name> GROUP BY datasliceid 9. What is collocated join? When you join tables which are distributed on the same key and used these key co lumns in the join condition, then each SPU in netezza works 100% independent of the other, as the required data is available in itself. This type of joins is ca lled collocated joins. 10. When netezza redistributes a table and when it broadcasts a table? Whenever it is not possible to do a collocated join, netezza either redistribute s the tables or broadcasts the table. When the table is a small one, then netezz a broadcasts the table. Otherwise netezza redistributes the table. 11. How do you remove logically deleted records? Whenever you delete a row in a table, it is not physically deleted. It is logica lly deleted by flagging the deletexid field in the table. NZRECLAIM utility is u sed to remove the logically deleted records. 12. What is nzload? Nzload utility is used load data from a file into a table. It is used to load bu lk data quickly and simultaneously rejects erroneous content. 13. What are the ways to load a data from a table into a file? Create an external table. Use nzsql utility with -o option. 14. What are the different ways to load data from a file into a table? Use nzload to load the data from a file into a table Create an external table and then load the original table using the external tab le. 15. How netezza updates a row in a table? Netezza logically deletes the original row by flagging the deletexid column with the current transaction id and inserts a new row with the updated values. SELECT datasliceid,* FROm lu_str_gmg_metric GROUP BY datasliceid select version(); 'Release 4.6.8 (P-7) [Build 15777]'

Das könnte Ihnen auch gefallen