Talk:OLAP cube

PK Preserving?
What does the comment "hopefully PK-preserving" mean? Please explain. --MauriceKA 09:36, 17 Mar 2005 (UTC)
 * I do not know the answer to that, but I'd like to point out that Google has only one page of results for "pk-preserving," the first result being this article, with the rest being mirrors. Can someone knowledgable explain if this is even a real concept? In the meantime, I have placed a cleanup tag on it for being technical. --Migs 08:46, 8 August 2005 (UTC)
 * PK in that context almost certainly refers to Primary Key. Unfortunately, I am insufficiently knowledgeable about Database Theory to cleanup the article myself. --Eyrian 18 August 2005
 * I believe however put it meant that the Primary Key must be retrieved somehow in the OLAP cube projection. Anyway, Im taking it off.  SSPecter talk &spades;  08:06, 6 January 2007 (UTC).

Preserving privacy while using authorization certificates ?

Other cubes
Shouldn't this article reference ROLAP, MOLAP, and HOLAP cubes as well? 65.202.85.9 15:23, 11 January 2007 (UTC)
 * I don't think so. ROLAP, MOLAP and HOLAP are all about different physical storage methods. The logical essense of OLAP Cube is the same regardless of physical storage details. Wikiolap 17:03, 11 January 2007 (UTC)


 * I think it should relate to all the other OLAPs, since a cube is a MOLAP and others like ROLAP (relational OLAP) are totally different concepts and ways of preparing the data to be queried. —Preceding unsigned comment added by 12.145.46.162 (talk) 23:36, 11 September 2007 (UTC)

Functional notation
Can someone please explain the functional notation here? I believe it should be f: (X, Y, Z) |--> W, not W: (X, Y, Z) |--> W. Ericdbw 06:15, 16 September 2007 (UTC)

I agree. I've changed it. Qseep (talk) 00:08, 10 February 2011 (UTC)

Functionality
This needs to be included in the article:

The OLAP cube consists of numeric facts called measures which are categorized by dimensions. The cube metadata is typically created from a star schema or snowflake schema of tables in a relational database. Measures are derived from the records in the fact table and dimensions are derived from the dimension tables.Kgrr (talk) 11:55, 18 March 2008 (UTC)
 * I have a couple problems with this bit
 * It is mis-captioned. Functionality - "That set of functions that something is able or equipped to perform" (Wiktionary) is not addressed in this section.
 * Without an authoritative citation, I am not sure how "typical" the star schema formulation is. While it may be typical of particular contexts, I doubt it is typical of OLAP in general. anecdotally, I have seen far more installations that bear no relation to a star schema than ones that do.
 * Perhaps one could expand further on these ideas in a better context. Ningauble (talk) 17:56, 30 July 2008 (UTC) (reformat comment Ningauble (talk) 16:09, 31 July 2008 (UTC))

Timestamp Example
The timestamp example does not clearly explain the benefit of assigning an ID value to a timestamp when the timestamp itself could be used. It seems like useless abstraction. Could someone more knowledgeable provide a more clear example? —Preceding unsigned comment added by 24.165.146.46 (talk) 02:32, 13 June 2009 (UTC)

OLAP Cube (introductory) section
Text says: An OLAP (Online analytical processing) cube is a data structure that allows fast analysis of data. It can also be defined as the capability of manipulating and analyzing data from multiple perspectives. This is followed by a reference to an article by Codd. I have the following issues with the above:
 * I don't think an OLAP cube is a recognized data structure, but is rather, perhaps, a conceptual data model.
 * the referenced article by Codd does not contain the words cube, fast, or analysis.
 * the second sentence (It can also be defined as...) is incorrect. An "OLAP cube" is not a "capability".

The section goes on to say that "Relational databases are not well suited for near instantaneous analysis and display of large amounts of data", despite the existence of ROLAP and the fact that some proponents of the relational model think performance is a red-herring (after factoring in the need for integrity constraints processing outside of the DBMS). For example, see http://www.information-management.com/issues/20020601/5251-1.html. So I don't think this sentence is well founded. At a minimum it needs a reference and I have marked it up as dubious/citation needed.

The following sentence is presented as a statement of fact, but does not provide any references: "Although many report-writing tools exist for relational databases, these are slow when the whole database must be summarized." The problems I have with this statement are:
 * "slow" is meaningless without context; what is slow for some application may be perfectly acceptable in other uses. Since the intention of this section is to make the case for OLAP cubes, I think this should say "slower than using OLAP cubes", or some such.
 * a good reference should be cited that is accepted by the mainstream DBMS community as authoritative. It would have to compare like for like results showing where OLAP cube performance is typically better than "report writing tools for relational databases", whatever that is  —Preceding unsigned comment added by 47.166.133.98 (talk) 16:48, 19 October 2009 (UTC)

I think Microsoft MSDN page provides a better explanation to what an OLAP cube is. I have attached the link here. I am not sure if we could use the information from this link to update the Introduction. Database experts can please verify this. Thank you. Link : Microsoft MSDN Article -- Abdul Jabbar 09:27, 28 April 2010 (UTC) (talk)

NPOV dispute
Language like

"The arrangement of data into cubes overcomes a limitation of relational databases. Relational databases are not well suited for near instantaneous analysis and display of large amounts of data."

sounds like an advertisement, not NPOV. Crasshopper (talk) 05:05, 2 March 2010 (UTC)

Sounds factual to me, doesn't a cube's creation implicitly preprocesses a lot of the possible selections therefore greatly reducing selection time at a later time? So if the data is maintained in a cube format, selections from it will be fast.

131.111.202.114 (talk) 13:14, 13 March 2010 (UTC)


 * A cube can be implemented in rdb, so where's the point ? The difference that could have been emphasized is transaction-optimized vs read-mostly optimized, instead. The current wording is not factual and rather misleading marketoid talk. --Lysytalk 22:26, 21 March 2010 (UTC)

As a newbie to OLAP, when I read the sentence about RD not being suited for instantaneous analysis, my first question was "so how do you implement it". Later on in the article, it says that you implement it in a relational database... The claim needs at least some explanation, where it stands in the text. I can't disagree completely with the marketing talk claim, since I have implemented relational models on non-relational flat file databases using tricks, and I wouldn't say that this fact makes relational databases the same as, say a flat file database. The fact that an OLAP cube CAN be implemented on a relational database does not make it an analogous concept to a relational database. From the abstract description of the OLAP cube, it is not clear that it must be implemented on a relational database; and I'm guessing that in fact it can be implemented otherwise. 62.233.239.26 (talk) 06:20, 7 May 2010 (UTC)J

It sounds to me like OLAP Cubes are an abstract data concept. It may be that certain implementations are designed to be efficient at certain actions but I'd want to see citations (from something other than product marketing material). Also, reference #4 -- cited to back-up claims about efficiency of OLAP vs RDBMS -- is a link to an answers.com page which is, at least now, almost entirely just a copy of the wikipedia doc. I see nothing there of relevance. 83.67.12.45 (talk) 07:52, 16 May 2010 (UTC)

From what I know, OLAP would use Star and Snowflake schema. These would be much less normalized than schemas use in Operational Relational Databases. They would not enforce/check for constraints such as uniqueness, foriegn key references etc making them faster than Operational Relational Databases even if they are implemented using RDBs

nishantjr (talk) 10:42, 15 June 2010 (UTC)

Disambiguation of "residual"
This article currently links to residual, which is a disambiguation page. The link needs to get replaced by residual. Michael Hardy (talk) 19:59, 5 May 2011 (UTC)

Non-English text inside images
Looking at the images, they contain text that is not in English. Looks like German to me, or some nordic language perhaps. Can anybody translate them? Thanks. The 11th plague of Egypt (talk) 15:42, 21 March 2014 (UTC)
 * I did add a translation request but it was reverted with "you need to take this to Graphic Lab". I had been to the graphics lab before listing it for translation here, and did not find anything translation related there. Jay (talk) 19:18, 29 April 2021 (UTC)
 * I would imagine one of the workshop pages linked there, probably Graphic Lab/Illustration workshop based on the instructions. There is nothing we can do about it here or at PNT as they are not hosted here.-- Jac 16888 Talk 14:44, 30 April 2021 (UTC)
 * I have put a question on Commons:Commons talk:Graphic Lab/Illustration workshop though I don't see why a graphics lab would be interested in translation. What I am interested in is the translation, not the graphics. I can do the graphics. Jay (talk) 06:48, 14 May 2021 (UTC)
 * Commons have done this before when I have referred other editors, PNT is for articles since we don't have the capability for editing images. The terms themselves however are:
 * Camping equipment
 * Accessories
 * Outdoor protective equipment
 * Golf Equipment
 * Mountaineering equipment
 * Northern Europe
 * Central Europe
 * Southern Europe
 * Insect repellent
 * Sun Block
 * First Aid
 * Going down the first column, then across the bottom row, then down the second column-- Jac 16888 Talk 16:21, 14 May 2021 (UTC)
 * Thanks for the translation. This works for me. I'll keep an eye on Commons. Jay (talk) 18:40, 22 May 2021 (UTC)
 * I have created the English images and replaced them in the article. Also, I found the page in commons where the text translation can be done (Commons:Requests for translation), following which it has to be listed at the Graphic Lab, so it is a two-step process. Jay (talk) 21:55, 23 May 2021 (UTC)

Mathematical definition
The unmathematical definitions on this article confuse me. In the "Mathematical definition" section, f is not defined, so it might be anything including the identity function (W is not defined either). So is the cube just an Rn or Xn space, or a multidimensional array? --Nemo 14:13, 9 June 2015 (UTC)

Seems to me we're really talking about hyperrectangles, not hypercubes, but if the industry has accepted 'hypercube', we're too late. Not the first time I've seen this conflation. Wootery (talk) 17:21, 4 February 2018 (UTC)