|
OMNIDEX |
Partitioning |
|
Step 1 - AnalysisThe first and most important step in partitioning a table is analysis to determine if the table should be partitioned, and if so, how. Analyze the queries that will run against the partitioned table to check for SQL limitations and to determine the best partitioning approach.
The Omnidex utility, ODXSQL, can assist in determining the best partitioning approach by simulating the partition in a variety of ways and generating a report detailing exactly how the data would be partitioned. When you've completed your analysis, ODXSQL can then physically partition the table for you and generate the necessary environment file entries to represent the new partitions. See Step 2 - Partition the Table, for more information on using the ODXSQL PARTITION command.
Partition QualifierA partition qualifier is the value on which the partition will be based. For example, if the table is partitioned by state, the partition qualifier for each partition will be the state or states stored in that particular partition. The partition qualifier field should be a reasonably low-cardinality field, meaning the number of unique values in that field for each partition should be low. The field used for the partition qualifier must be a reasonably low cardinality field. When the partitions are created, each individual partition should have no more than 256 unique values in the partition qualifier field. This is important for performance. For example, partitioning a customers table based on social security number is not very efficient because every social security number is unique. A social security number field has very high cardinality. However, partitioning the same table based on state makes more sense since there are only 50 states. |
|
Environment Catalog | Database Integration | Omnidex Text | Other Features |
Environment |
Database |
Table |
Column |
Rule |
Index |
Message |
SQL Server |
Oracle |
DB2 |
Flat Files |
External Documents |
Configuration |
Keyword Searches |
Synonym Searches |
Proximity Searches |
Form Searches |
Misspelling Searches |
Spell Check Searches |
Phonetic Searches |
Exporting Data |
Pseudo-columns |
Partitioning |