Talend ETL - Open studio version

A lovely opportunity landed in my way to try Talend Open studio. It was a breeze to adopt Talend Open studio. In specific, I used 4.0.3 version. The highlight is, it is built on top of eclipse platform, enabling any developer with fundamental knowledge of Java language to be easily extra productive. I cannot comment by comparing this with commercial tools in the market as I haven't used other tools before. One aspect I am sure is... the arrival of open source tools in the BI space will soon change the BI solution offerings are made. I am convinced Talend as a basic ETL tool and not sure if this can be used for "Heavy lifting".

Technology radar - Jan 2011 is available

http://tinyurl.com/2em9mnc contains the technology radar published in Jan 2011. Interesting !

Installing Oracle 11gR2 - Linux x86 - owb/external/oc4j_applications/applications/WFMLRSVCApp.ear not found

Installing Oracle 11gR2 on Linux x86.  It is deemed a simple task as long as we unlearn the way we install previous versions of Oracle and pick this up.

Both Disk1 and Disk2 should be unzipped in to same location .. what I mean is.. unzipping both will result in single directory named database. Launch installer from this location.

Otherwise, resolving kernel parameters and package dependencies are rather simple.

MySQL has the real syntax sugar

My first experience with MySQL is so wonderful. I never imagined that profiling queries can be this easy. Oracle profiling has a learning curve. SQL server is also somewhat simple and straightforward. Looking forward to see how it is implemented in PostgreSQL.

A little odd thing I see is the choice of storage engines. Otherwise, MySQL journey has been very smooth. One thing is sure, all relational databases architecture is kind of similar. Learn internals of one rdbms.. and believe me.. it is really  fun and adapt to work on others.

Calculating size of a table in Oracle

Before calculating the size of a table, one should understand how a table is stored.

A table is physically stored in a datafile on disk. Oracle reserves a defined space before writing the data. The data file is organized in to chunks called segments, which further contains chunks named extents and each extent is organized in to most atomic units called the blocks. The size of the block is determined by the Oracle init parameter named "db_block_size".


The data dictionary stores the details of filled and free blocks or extents. (You should do a step called "analyze" to gather the table statistics)
something like analyze table MYTAB compute statistics;
This might take long time depending on your tablesize.

And now you are ready to query the data dictionary tables to find the size occupied by the table.

select sum(blocks) from user_extents
where segment_name = 'TABLE_NAME'
and segment_type = 'TABLE';

or 

select blocks,EMPTY_BLOCKS,avg_space,avg_row_len,NUM_FREELIST_BLOCKS  from user_tables
where table_name='TABLE_NAME';

My personal choice is the first query.

Final steps ...
1. Know the block size by running "show parameter db_block_size" in sqlplus.
2. Multiply the block size by sum(blocks), and you get the table size.

SQL Server 2008 for the Oracle DBA

For those who have lived long enough working Oracle, and if you wish to learn SQL server 2008 by mapping feature set to Oracle database, then this is the video for you.

http://technet.microsoft.com/en-us/sqlserver/dd548020.aspx

Quick Summary from the link:
This 15 modules, level 300 course provides students with the knowledge and skills to capitalize on their skills and experience as an Oracle DBA to manage a Microsoft SQL Server 2008 system. This workshop provides a quick start for the Oracle DBA to map, compare, and contrast the realm of Oracle database management to SQL Server database management.