10,382 reputation
11336
bio website
location Munich
age
visits member for 2 years, 11 months
seen 50 mins ago

I'm looking for job as a software developer (Java) with focus on relational database technologies (Oracle, PostgreSQL)


Mar
14
comment CentOS and Ubuntu differences with information_schema
I know it doesn't - but what is the use of accepting a statement without an error message that doesn't work? If you are migration existing scripts for other DBMS you might think everything is OK because MySQL did no complain (I have seen more than once) - it's the same story with check constraint.
Mar
13
comment Compare Number of Rows of Tables
@GismoRanas: it's not "Post", "gre" and "SQL" it's "Postgres" and "SQL" and the duplicated s is dropped. I think they should simply consider writing it differently: "PostgresQL", instead of "PostgreSQL" (because they can't really change the name any more)
Mar
13
comment CentOS and Ubuntu differences with information_schema
@jynus: there are still situations where foreign keys are silently ignored: sqlfiddle.com/#!2/db0d8/1 (try that with any other DBMS...)
Mar
12
comment CentOS and Ubuntu differences with information_schema
@watcher: or use a DBMS that fails if it can't comply with a request rather than trying to be smart and do "something" just to avoid an error message (that situation can still happen with current MySQL versions if for some reason the InnoDB engine could not be loaded/started)
Mar
12
comment Sharding Postgres Database
Please define "fairly large". Unless each and every table goes into hundreds of million of rows (and you have like hundreds of those tables) I don't think adding the complexity of sharding is worth the effort.
Mar
12
comment CentOS and Ubuntu differences with information_schema
I would say: whatever script you use to install/create your tables, that script didn't run properly on the CentOS machine. Another possibility: you don't have InnoDB enabled on CentOS and therefor your foreign keys weren't created.
Mar
12
comment Sharding Postgres Database
I wonder how big your customer table is. We are happily using tables with over 50 million rows. You must have some really good running business if you have more than 50 million customers. If you really, really need this you might want to take a look at Postgres-XC
Mar
12
comment mysql id sequencial problem
Quick answer: you don't. The generated values of a primary key have no meaning whatsoever. If you rely on gapless numbers in your primary key, your design is seriously flawed.
Mar
11
comment Use of integer instead of interval (of one type)
@vedic: I didn't really think that an int2 will be slower than an int8 on processor level. My point was that it won't be a difference once the data is in memory.
Mar
10
comment How to print the information of all tables using Oracle?
What exactly is content1 to be? If you want to show the content of the tables you potentially have may rows for each table.
Mar
10
comment Use of integer instead of interval (of one type)
Plus I always think that modern 64bit processors can't handle a two-byte integer anyway. I might be mistaken, but I don't think they can physically handle less than 64bit, everything else is just setting all the other bits to zero (but I could be mistaken there)
Mar
9
comment What does: ERROR: tuple offset out of range: 0 mean?
Does rebuilding the index (or even all indexes on the table) help?
Mar
9
comment How large should a database table grow?
How many rows does the table have? And what is the definition of the table? What kind of columns does it contain?
Mar
8
comment Use of integer instead of interval (of one type)
Sounds OK to me.
Mar
6
comment What is the difference between MySQL and PostgreSQL in terms of writing SQL queries?
Other not so obvious things: you can't create a view in MySQL with a derived table in the select and you can't use dynamic SQL in functions.
Mar
5
comment What is the difference between MySQL and PostgreSQL in terms of writing SQL queries?
Postgres has much more advanced SQL features than MySQL (CTEs, Window functions, recursive queries, deferrable constraints and many more), here is an overview sql-workbench.net/dbms_comparison.html
Mar
5
comment Can someone explain why left joining two views in mysql is so slow?
Does this mean MySQL is not able to use indexes on the underlying table when querying a view?
Mar
4
comment Delete millions of rows, fast
In my experience creating a new table with only the data you want to keep is usually much faster than deleting large parts of a table. Of course this only works if your can afford to drop the real table at some point (so the application accessing it needs to be stopped as well most probably)
Mar
4
comment 25M x 25M inner join (postgresql) performance
Indexing won't really help as you do want all the rows from both tables. There is no index that can speed up this query.
Mar
4
comment Handling concurrent updates, inserting if and only if no record already exists
I have no idea if advisory locks could be simulated in H2. The reason why I think SERIALIZABLE works better in Postgres, is that I am not sure that H2 implemented it as efficiently as Postgres: wiki.postgresql.org/wiki/Serializable and drkp.net/papers/ssi-vldb12-slides.pdf