Interesting observation on MySQL

[ichabod801]

I shudder when I hear the phrase "sql server," but honestly that has more to do with the horrid database I have to deal with in sql server. The server and the sql don't give me problems, the database does, but they're too closely associated in my mind.

[nilamo]

I didn't say I *like* sql server, it's just what I deal with these days lol. That said, the Expected/Actual Execution Plan is a nice way to make sure I'm not doing a terrible job, or to stumble across something I thought already had an index.

Speaking of indexes, it seems to need them rebuilt far more often than I'd like.

[Larz60+]

SQL server is still much more capable than MySQL

Relational databases make connecting data together much easier, but ...

I expect that some of you have had to deal with capturing packets (phone call packets for example, before wireless was commonplace)
that were flying by your switch back in the early 90's, at at unheard of (at the time) pace.
All Relation databases were too slow to deal with the volumes of transactions that had to take place.
Up to the point where call data could be summarized into billable units, everything had to be dealt with using hashed flat  files (whether in memory
or on disk) which were essentially what Python calls dictionaries.
After that point, the RDBMS took over

The database was one of Oracle, Sybase, Informix or homebrew.

Of course, PostgresSQL didn't exist (it actually did, but Berekley only released a few copies to very brave volunteers back in June of 1989)
I wasn't offered, and I am sure wouldn't have been allowed to use it even if I had been.

MySql wasn't released until May of 1995.

PostgreSQL is now the most advanced open-source database available anywhere.
I bit difficult to use, with the steep learning curve, but rich in performance and toolset.

[wavic]

Hm! Why I am reading all over the web that relational databases cannot meet today's requirements?

[Larz60+]

SQL server is light years ahead of MySQL
also of note (about PostgreSQL:

It’s a cousin of the Sybase and Microsoft SQL Server databases, because the
people who started Sybase came from UC Berkeley and worked on the Ingres or
PostgreSQL projects with Michael Stonebraker. Michael Stonebraker is considered by
many to be the father of Ingres and PostgreSQL and to be one of the founding fathers
of object-relational database management systems

I forgot about Ingres, and did use it at one point. Don't remember anything good or bad about it though.

[Larz60+]

Hm! Why I am reading all over the web that relational databases cannot meet today's requirements?

My last post actually supports the fact that relational databases aren't the answer to everything. By the
nature of the complexity of the software behind one, they will unlikely ever surpass a random access flat
file indexed by a hash table.

That doesn't exclude their usefulness. Just like you wouldn't use a jack hammer to put a 16 penny nail into
a 2 x 6, it still has it's use.

If you study the internet, You also find plenty of claims that Microsoft windows 10 or whatever version
you'd like it the best thing since sliced bread, even though it still is not as robust as the first Unix I used
at Bell Labs back in the early 80's.

In my opinion, windows is not the best OS, but it is the most popular., and it has it's use.

final note, when someone writes: relational databases cannot meet today's requirements, they should
list what those requirement are. Seems to me I heard the same argument back when C. J. Date came
out with SQL/DS and DB2.

It's all relative

[ichabod801]

People talking about SQL being obsolete seem to be talking about big data. However, I'm not really impressed with big data, and most people don't have big data, so I think SQL still has a lot of uses. As Larz60+ demonstrates, SQL can also have problems with fast data, even if it doesn't get that big. But again, not everybody has or needs that kind of data.

Management at work is constantly asking us about being ready to get our big data on the cloud. We keep telling them we don't have big data. Our biggest data set would be NEISS, the usable part of which is maybe seven or eight million rows. That's not big data.

[Ofnuts]

Cloud providers want you big data because once they have them they become extremely difficult and costly to move (and if you read the EULA, if they lose them they will just say "oops, sorry!" even if that puts your company out of business).

[Larz60+]

Cloud is a very bad idea. Ofnuts is absolutely correct. You should read Woz on the cloud: http://www.businessinsider.com/steve-woz...ars-2012-8

Seems as though the Patent office would qualify as Big Data (I don't like this term, but I guess it's here for a while)

Management should manage, and keep out of decision making.

Exception - If the manager grew up from a technical position. (this was me)

[ichabod801]

Cloud is a very bad idea. Ofnuts is absolutely correct.

Believe me, I am not arguing with either of you. Especially because our data contains confidential medical, personal, and business information.