Talk:Cardinality (SQL statements)

Latest comment: 14 years ago by GregorB in topic Low cardinality and use of Index.
WikiProject iconDatabases Start‑class (inactive)
WikiProject iconThis article is within the scope of WikiProject Databases, a project which is currently considered to be inactive.
StartThis article has been rated as Start-class on Wikipedia's content assessment scale.

Cardinality in math edit

Shouldn't this article point to Cardinality, somehow?
--Jerome Potts (talk) 05:30, 11 January 2008 (UTC)Reply

SQL edit

Cardinality is not peculiar to SQL. GregorB (talk) 20:02, 30 January 2008 (UTC)Reply

Input from DerekAsirvadem edit

I removed the following paragraphs from the 05:42, 29 January 2009 version of the article, as the notes are more appropriate on the discussion page. The notes were added by DerekAsirvadem. Currently, I don't have an opinion about the notes. But let's discuss them.

1 There is no such thing as "normal" cardinality. In order for such a concept to exist, there must be some pre-existing measure (there is not). Cardinality is a relative term, relative to the object: if the table has a million rows, and there are close to a milliion values, the column has high cardinality; if the values are few, the column has low cardinality. There is no measure outside the table against which "normal" can be determined.

2 The highest level of cardinality is uniqueness: one unique value per row. That may well be a range of more than one million values.

3 The definiions are adequate but the examples are poor.

Troels Arvin (talk) 09:17, 29 January 2009 (UTC)Reply

Low cardinality and use of Index. edit

I am using SAP and one mandatory field in most of the tables that is MANDT. This is taken a s primary key or a part or a secondary index key and becomes a pre included field in all the queries. Even if an ABAPER does not write it so, till it comes as a result of translation by SAP. Typically a production system does not have many client or rather say a few clients only. So this is a very low cardianlity column in most of the table consisting millions of rows...

There is also a rule which says that the indexing a column of low cardinality type to be avoided.. and hence there is the confusion.

What is correct? Indexing or not in a low cardianlity column?


59.160.30.29 (talk) 07:21, 25 March 2009 (UTC)Soumen(http://soumen.wordpress.com)Reply

OK, this is not what talk pages are for, but here it is anyway... Indexing a column of low cardinality makes sense when:
  • using bitmap indexes
  • the column in question is a part of a composite index. GregorB (talk) 19:32, 17 July 2009 (UTC)Reply