Category Archives: Performance

Strange behavior in MERGE with sequence NEXTVAL

Today we learned the following: “When using a sequence in a MERGE statement, the nextvalue is generated for all rows both updated and inserted (hence all rows in the source).” Our initial reaction was that this must be a bug, but is it really?

Redundant NOT NULLs in DDL

I often see DDLs with NOT NULL definitions which is really redundant. When I see such DDLs I really question “do the writer really know what he or she is doing”. They might not hurt you application, but why set them if they are redundant? (If you know I reason please comment) Example: CREATE TABLE […]

Java: JDBC roundtrips are expensive

SQL Daily just tweeted about the fact that JDBC roundtrips are expensive. This is something many developers either forget or do not know. I have seen so many chatty applications killing the application performance, and the database performance. Toon Koppelaars have performed some excellent presentations showing this. Lukas Eder have also written some nice articles […]

OOW16 – Are we failing or will performance rise?

The last session I attended at OpenWorld 2016 was “Thinking Clearly About Database Application Architecture”. A panel discussion with Toon Koppelaars, Cary Millsap, Gerald Venzl, Bryn Llewellyn and Connor Mcdonald. The panelists promotes that the business logic should be moved into the database, and implemented using PL/SQL. And yes – I do get the message, but I just […]

Books to read: “Mastering Oracle Trace Data, Second Edition” with Cary Millsap

The “Mastering Oracle Trace Data” from Cary Millsap is a great book about performance in general, and about how to work with trace data. The second edition have several new chapters adding valuable knowledge to both developers and DBAs. This is one of the books that you really should have in your bookshelf.

One of the things you just need to know about Oracle …

A couple of years ago I did ask my developers to start using the varchar2(char) option when creating new tables. In Norway we have some characters in the alphabet (‘Æ’, ‘Ø’, ‘Å’) that uses more than 1 byte of storage (when the database characterset is set to AL32UTF8). Using the default – varchar2(byte) – could then decrease the number […]

Old blog from Antognini regarding trace levels (10046)

Link to the blog: http://antognini.ch/2012/08/event-10046-full-list-of-levels/

Oracle 11g: New Locking behavior (modes) … an unexpected behavior. Or?

I just noticed some strange locking behavior in one of our production databases. It took me a while to reproduce this in test. I’ve based my example on Richard Foots blog “Oracle11g: New Locking Modes When Policing FK Constraints (A Wolf at the Door)”.