- Determine the cardinality of column combinations. This is often an easy way to answer the question as to which the adding of another column to the index actually increases the selectivity of the index.
- Display unequal distributions in the column values. It is displayed how many values there are for the individual column values or combinations of column values.
 | Warning This procedure is not suitable for large tables because the entire table is read. Depending on the size of the table this is very expensive. |
In transaction DB05 you enter the table name and description of the columns to be analyzed.
In the upper part, the result returns the table size - this value is only correct if the primary key columns have been selected - (Requested bytes to be buffered 100%), the number of records of the table (Total number of rows) and the SAP buffer status (Current buffering mode).
The bottom table shows how many distinct values the combinations of analyzed columns have and how many records have to be read for the different combinations. Distinct values: Number of different values for an index (in the example below there are two values for the client and 418 different values for the combination of client and name_last. Number of areas that are specified by the generic key and contain the given number of rows:
Number of records that would have to be read for a certain value combination of index columns (in our example 1 to 10 records would have to be read for 415 of the value combinations, but there is also one value combination each for which 11 to 100, 101 to 1000, and 1001 to 10000 records would have to be read).
If an additional column is added to the index, you can check the resulting selectivity. In our example, adding the birthdt column to the index does not increase the selectivity! Extending the index would not result in an improvement.
No comments:
Post a Comment