Home
AdministeringIn addition to administering the database server, you can tune performance, replicate data, and archive data.
System administrationThese topics contain concepts, procedures, and reference information for database and database server administrators to use for managing and tuning HCL OneDB™ database servers.
Performance GuideThe HCL OneDB™ Performance Guide describes how to configure and operate your HCL OneDB™ database server to improve overall system throughput and to improve the performance of SQL queries.
Queries and the query optimizerThese topics describe query plans, explain how the database server manages query optimization, and discuss factors that you can use to influence the query plan. These topics also describe performance considerations for SPL routines, the UDR cache, and triggers.
The query planThe query optimizer evaluates the different ways in which a query might be performed and determines the best way to select the requested data. During this evaluation, the optimizer formulates a query plan to fetch the data rows that are required to process a query.
The join planWhen a query contains more than one table, OneDB joins the tables using filters in the query. The way that the optimizer chooses to join the tables is the join plan.
Hash joinThe optimizer usually uses a hash join when at least one of the two join tables does not have an index on the join column or when the database server must read a large number of rows from both tables. No index and no sorting is required when the database server performs a hash join.

Hash join

The optimizer usually uses a hash join when at least one of the two join tables does not have an index on the join column or when the database server must read a large number of rows from both tables. No index and no sorting is required when the database server performs a hash join.

A hash join consists of two activities: first building the hash table (build phase) and then probing the hash table (probe phase). How a hash join is executed shows the hash join in detail.

In the build phase, the database server reads one table and, after it applies any filters, creates a hash table. Think of a hash table conceptually as a series of buckets, each with an address that is derived from the key value by applying a hash function. The database server does not sort keys in a particular hash bucket.

Smaller hash tables can fit in the virtual portion of database server shared memory. The database server stores larger hash files on disk in the dbspace specified by the DBSPACETEMP configuration parameter or the DBSPACETEMP environment variable.

In the probe phase, the database server reads the other table in the join and applies any filters. For each row that satisfies the filters on the table, the database server applies the hash function on the key and probes the hash table to find a match.

This figure is described in the surrounding text. — Figure 1: How a hash join is executed